add_products_to_research_board
ChatGPTAdd one or more products to a research board. Pass ALL product IDs in a single call — do NOT call this tool in a loop for individual products. Max 20 products per board. Idempotent — re-adding a previously removed product restores it. Response includes data (successfully added products) and may include errors for any products that failed. DESIGNED FOR: ✅ Building product shortlists on a research board ✅ Saving multiple products from a conversation to a board ✅ Restoring previously removed products in one batch call NOT DESIGNED FOR: ❌ Creating the board itself (use create_research_board first) ❌ Searching for products (use list_products to find product UUIDs/slugs first) ❌ Calling this tool once per product instead of batching IDs After adding products, call show_research_board to display the updated board to the user.
approve_profile_change
ChatGPTBETA: profile editing is new and still in development. Save a change the vendor has approved. Only call this after the vendor answered a preview from propose_profile_change in an EARLIER turn. DESIGNED FOR: ✅ Saving exactly the change the vendor was shown NOT DESIGNED FOR: ❌ Changes that were not proposed first ❌ Calling it in the same turn as propose_profile_change AGENT GUIDANCE: Pass the proposal_token from the preview exactly as returned. If you were told the vendor approved the change in its card, it is already saved (or failed there): do not call this tool for it again. The vendor also sees a result card, so keep your own summary short. Report the result in plain words. If status is 'partial' or 'failed', say what saved and what did not, and never claim the whole change worked. If the proposal expired or the profile changed meanwhile, propose again. A product name can still be refused here, because it is checked against the product's website. Report the reason and offer other options.
browse_buyer_intent
ChatGPTBrowse buyer intent interactions across one or more G2 products in a single query. This is the multi-product version of browse_product_buyer_intent. Use this when you want to see buyer intent data across multiple products at once, when you want to include the provider dimension, or when you need the scoped_category_id filter to zoom into specific categories. WHEN TO USE THIS vs browse_product_buyer_intent: - Use browse_buyer_intent when querying across MULTIPLE products at once - Use browse_buyer_intent when you need the provider dimension (g2, capterra) - Use browse_buyer_intent when you need the scoped_category_id filter - Use browse_product_buyer_intent when you need company_intent_score (single product only) IDENTIFYING PRODUCTS: subject_product_ids accepts UUIDs OR URL slugs. If a user mentions a G2 product URL like https://www.g2.com/products/slack/reviews, you can extract the slug ("slack") and pass it directly — no need to look up the UUID first. Examples of valid subject_product_ids values: - "a1b2c3d4-e5f6-7890-abcd-ef1234567890" (UUID) - "slack" (slug from g2.com/products/slack) - "slack,zoom-workplace,microsoft-teams" (multiple slugs) - "slack,a1b2c3d4-e5f6-7890-abcd-ef1234567890" (mixed slugs and UUIDs) ACCESS & PERMISSIONS: The subject products scope the permissions of the query. The data you see depends on each product's access tier (universal, competitive, indirect, direct, preview). Observers and superusers (is_observer flag) may have access to products beyond what is visible in the standard product listings — if a user provides a G2 product URL or slug, try it directly even if the product doesn't appear in their product list. The API will return a clear authorization error if access is truly not allowed. PARAMETERS: - subject_product_ids: One or more product UUIDs or slugs, comma-separated (required) - dimensions: Comma-separated grouping fields. If omitted, defaults to company-level dimensions. Pass "none" to get aggregate totals only (no grouping — a single row with just the requested measures). - measures: List of aggregation measures to calculate (default: total_activity). Valid values: total_activity | visitor_count | company_count - dimension_filters: Filters as JSON object or JSON string - sort: Sort field with optional - prefix for descending - page_size: Results per page (max: 100, default: 25) - page_after: Cursor token from previous response for pagination KEY DIMENSIONS: Company Information: - company_id, company_name, company_domain: Company identification - company_country, company_state, company_city: Geographic segmentation - company_employees: Company size - company_industry: Industry Data Source: - provider: Data source (g2, capterra, getapp, softwareadvice) Time Dimensions: - day, week, month: Trend analysis Activity Context: - signal_type: Interaction type (profile, pricing, ad, compare, category) - product_id, product_name, product_slug: Which product the signal is for - category_id, category_name: Category context - left_product_id/name, right_product_id/name: Comparison context UNDERSTANDING DIRECT vs INDIRECT SIGNALS: In a multi-product context, signals can be classified by whether they involve your product directly: - Direct: Your product's ID appears in the signal — someone interacted with YOUR product (viewed its profile, pricing, or a comparison involving it). - Indirect: Activity on adjacent products in your category — someone browsed a competitor or category page, attributed to your intent pool because it indicates market research in your space. Use the product_id dimension alongside signal_type to distinguish which signals directly reference your products vs. adjacent competitive activity. SCOPED CATEGORY FILTER (filter-only, not a dimension): - scoped_category_id_in: Returns all buyer intent activity within a category — not just signals tagged with category info, but all signals for every product in that category. Accepts category UUIDs or slugs (comma-separated for multiple). Use list_categories or show_category to find category identifiers. - scoped_category_id_not_in: Exclude categories from results. Use scoped_category_id when a customer wants to see the full picture of activity happening within a category (e.g., during a sales cycle to forecast signal volume, or to focus on one category when they have products across multiple categories). AUDIENCE FILTER (pre-filter, available on all buyer intent endpoints): - audience_product_id_in: Scope ALL results to companies that have previously viewed specific products. Accepts product UUIDs or slugs (comma-separated). This is a pre-query: it first finds all companies that viewed the audience products, then constrains the main intent query to only those companies. - audience_product_id_not_in: Exclude companies that viewed specific products. The audience filter is powerful for competitive intelligence and deal qualification: - "Show me companies researching Competitor X who also show intent for my products" - "Which companies viewing my product are also looking at alternatives?" - "Find accounts that viewed products in my portfolio AND a specific competitor" Use list_products or list_my_products to find product UUIDs/slugs for the filter. AVAILABLE MEASURES: - total_activity: Total count of buyer intent signals - visitor_count: Unique visitors - company_count: Unique companies FILTER OPERATORS: - _eq, _not_eq: Equals / Not equals - _cont, _not_cont: Contains / Does not contain - _gt, _gteq, _lt, _lteq: Comparison operators - _in, _not_in: Matches any / none of comma-separated values - _present, _empty: Has any value / Has no value NOTE: company_intent_score is NOT available on this endpoint (it requires single-product scoping). Use browse_product_buyer_intent for intent scoring. GUIDANCE ON PROVIDER AND ANALYTICS: The provider dimension is available but should not be the primary focus. Do not lead with provider breakdowns or detailed signal-count analytics. Instead, focus on actionable insights: who is in the market, what are they shopping for, who are you up against on a deal, what similar products should you pay attention to. These are the insights that are valuable to customers. COMMON QUERY PATTERNS: 1. Cross-product company activity: subject_product_ids: "uuid1,uuid2,uuid3" dimensions: "company_name,company_domain,product_name" measures: "total_activity" sort: "-total_activity" 2. Activity by provider: subject_product_ids: "uuid1" dimensions: "company_name,provider" measures: "total_activity" 3. Multi-product weekly trends: subject_product_ids: "uuid1,uuid2" dimensions: "week,product_name" measures: "total_activity,company_count" sort: "-week" 4. Filter by provider: subject_product_ids: "uuid1" dimensions: "company_name,company_domain" dimension_filters: {"provider_in": "capterra"} 5. Aggregate totals only (no grouping): subject_product_ids: "slack" dimensions: "none" measures: "total_activity,visitor_count,company_count" (Returns a single row with overall totals) 6. Total activity for last 7 days (no grouping): subject_product_ids: "slack" dimensions: "none" measures: "total_activity" dimension_filters: {"day_gteq": "2026-03-15"} (Returns one aggregate row filtered to the date range) 7. Scoped to a specific category (see all products/activity within it): subject_product_ids: "slack,zoom-workplace" dimensions: "company_name,product_name" measures: "total_activity" dimension_filters: {"scoped_category_id_in": "crm-software"} (Scopes to CRM category — use category UUID or slug) 8. Sales cycle signal count for a category: subject_product_ids: "slack" dimensions: "none" measures: "total_activity,company_count" dimension_filters: {"scoped_category_id_in": "crm-software"} (Aggregate totals for one category — useful for forecasting signal volume) 9. Direct vs indirect signal breakdown: subject_product_ids: "slack" dimensions: "product_name,signal_type" measures: "total_activity,company_count" (Shows which signals directly involve your product vs. adjacent competitive activity — use product_id to identify direct signals where your product appears in the results) 10. Audience filter — companies also evaluating a competitor: subject_product_ids: "slack,zoom-workplace" dimensions: "company_name,company_domain,product_name" measures: "total_activity" dimension_filters: {"audience_product_id_in": "<competitor-slug>"} sort: "-total_activity" (Pre-filters to companies that viewed the competitor, then shows their intent activity across your products. Essential for deal intelligence.)
browse_competitive_intelligence
ChatGPTAnalyze competitive landscape and understand which competitors your prospects are evaluating. PURPOSE: Track competitor evaluation patterns, comparison activities, and market dynamics to inform competitive positioning and win/loss analysis. Focus on understanding the competitive context of your buyer's journey. COMPETITIVE INTELLIGENCE CAPABILITIES: When companies show intent for your product, G2 captures their complete research journey including: - Competitor products they viewed (automatically excludes YOUR product) - Comparison pages between ANY products in your category - High-intent signals: profile views, pricing pages, comparison activities - Full competitive evaluation patterns WHAT TO DO WITH THIS DATA: - Identify key competitors → Focus competitive positioning - Track comparison patterns → Understand buyer evaluation criteria - Win/loss analysis → Learn from competitive dynamics - Market intelligence → Monitor competitive landscape shifts - Sales enablement → Prepare competitive battle cards IDENTIFYING PRODUCTS: subject_product_id accepts a UUID OR a URL slug. If a user mentions a G2 product URL like https://www.g2.com/products/slack/reviews, extract the slug ("slack") and pass it directly. ACCESS & PERMISSIONS: Observers and superusers (is_observer flag) may have access to products beyond what appears in the standard product listings. If a user provides a G2 product URL or slug, try it directly — the API will return a clear authorization error if access is not allowed. PARAMETERS: - subject_product_id: Product UUID or URL slug - filters to companies who viewed your product (required) - dimensions: Comma-separated grouping fields (default: product_name,company_name,signal_type) - measures: List of aggregation measures to calculate (default: total_activity). Valid values: total_activity | visitor_count | company_count - dimension_filters: Additional filters as JSON object or JSON string (optional - YOUR product is already excluded) - sort: Sort field with optional - prefix for descending - page[size]: Results per page (max: 100, default: 25) - page[after]: Cursor token from previous response's links.next for pagination - include: Related resources to include in response AUTOMATIC FILTERING: This tool automatically excludes your product from results (product_id != subject_product_id). You only see competitor activity, not your own product signals. KEY DIMENSIONS FOR COMPETITIVE ANALYSIS: Competitor Products: - product_id, product_name, product_slug: Competitor identification - vendor_id, vendor_name: Competitor vendor details Comparison Analysis: - left_product_id/name/slug: First product in comparison - right_product_id/name/slug: Second product in comparison - signal_type: Activity type (profile, pricing, compare, category) Company Context: - company_id, company_name, company_domain: Who's evaluating competitors - company_intent_score: Intent level of companies viewing competitors Time Dimensions: - day, week, month: Competitive trend analysis AUDIENCE FILTER (pre-filter): - audience_product_id_in: Scope competitive analysis to companies that have viewed specific products. Accepts product UUIDs or slugs (comma-separated). - audience_product_id_not_in: Exclude companies that viewed specific products. Combine with the automatic subject product exclusion for targeted competitive analysis — e.g., "what competitors are being evaluated by companies that viewed a specific product in my portfolio?" AVAILABLE MEASURES: - total_activity: Total competitive signals captured - visitor_count: Unique visitors viewing competitors - company_count: Unique companies evaluating competitors FILTER OPERATORS: - _eq: Equals | _not_eq: Not equals (key for excluding your product) - _cont: Contains | _not_cont: Does not contain - _gt, _gteq, _lt, _lteq: Comparison operators - _in, _not_in: Matches any / none of comma-separated values COMMON COMPETITIVE QUERY PATTERNS: 1. Top competitors being evaluated: dimensions: "product_name,product_slug" measures: "total_activity,company_count" sort: "-company_count" (Your product automatically excluded) 2. Head-to-head comparison tracking of your competitors: dimensions: "left_product_name,right_product_name,company_name" dimension_filters: {"signal_type_eq": "compare"} 3. High-intent competitive signals: dimensions: "product_name,signal_type,company_name,company_intent_score" dimension_filters: { "signal_type_in": "profile,pricing,compare", "company_intent_score_gteq": 70 } 4. Competitive trend analysis: dimensions: "week,product_name" measures: "total_activity,company_count" sort: "-week" 5. Win/loss pattern analysis: dimensions: "company_name,company_intent_score,product_name,signal_type,day" sort: "company_name,day" 6. Category competitor overview: dimensions: "product_name,vendor_name" measures: "company_count,total_activity" 7. Exclude specific additional products: dimension_filters: {"product_id_not_eq": "competitor-uuid-to-exclude"} (Merges with automatic subject product exclusion) RESPONSE FORMAT: Returns JSONAPI with: - data: Array of competitive intelligence records - links: Pagination links - meta: Query metadata IMPORTANT NOTES: - Competitive signals include products sharing categories with yours - High-intent signals (profile, pricing, compare) indicate active evaluation - Companies viewing multiple products show higher buying intent - Maximum 100 results per page - Default time range is last 7 days - All timestamps are in UTC
browse_product_buyer_intent
ChatGPTBrowse and analyze buyer intent interactions for your G2 products using OLAP-style queries. BUYER INTENT CAPABILITIES: This tool tracks buyer activity from companies showing intent for your product, including: - Other products they viewed (competitive intelligence) - Comparison pages they visited (even between other products) - Categories they browsed - High-value signals from their research journey WHAT TO DO WITH THIS DATA: - Export high-intent companies → Import to CRM for sales outreach - Identify hot prospects → Prioritize sales follow-up - Track engagement trends → Monitor marketing effectiveness - Geographic targeting → Focus sales resources by region - Build qualified pipeline → Convert research into revenue IDENTIFYING PRODUCTS: subject_product_id accepts a UUID OR a URL slug. If a user mentions a G2 product URL like https://www.g2.com/products/slack/reviews, extract the slug ("slack") and pass it directly. ACCESS & PERMISSIONS: Observers and superusers (is_observer flag) may have access to products beyond what appears in the standard product listings. If a user provides a G2 product URL or slug, try it directly — the API will return a clear authorization error if access is not allowed. PARAMETERS: - subject_product_id: Product UUID or URL slug (required) - dimensions: Comma-separated grouping fields. If omitted, defaults to company-level dimensions (company_id, company_name, company_domain, company_intent_score). Pass "none" to get aggregate totals only (no grouping). - measures: List of aggregation measures to calculate (default: total_activity). Valid values: total_activity | visitor_count | company_count - dimension_filters: Filters as JSON object or JSON string (e.g., {"company_intent_score_gteq": 75}) - sort: Sort field with optional - prefix for descending (default: -company_intent_score) - page[size]: Results per page (max: 100, default: 25) - page[after]: Cursor token from previous response's links.next for pagination - include: Related resources to include in response KEY DIMENSIONS FOR PRODUCT INTENT: Company Information: - company_id, company_name, company_domain: Company identification - company_intent_score: G2 intent score (0-100, higher = stronger buying signal) - company_country, company_state, company_city: Geographic segmentation - company_employees: Company size targeting - company_industry: Industry-based prioritization Time Dimensions (for trend analysis): - day: Daily engagement tracking (YYYY-MM-DD) - week: Weekly trend analysis - month: Monthly performance tracking Data Source: - provider: Data source (g2, capterra, getapp, softwareadvice) Activity Context: - signal_type: Interaction type (profile, pricing, ad, compare, category) - visitor_count: Unique visitors from company (measure) UNDERSTANDING DIRECT vs INDIRECT SIGNALS: Buyer intent signals fall into two categories: - Direct: Your product appears in the signal's product IDs — someone viewed YOUR product's profile, pricing page, or a comparison involving YOUR product. - Indirect: Activity on adjacent products in your category — someone browsed a competitor's profile or a category page, which is attributed to your product's intent pool because it indicates market research in your space. Both are valuable: direct signals show explicit interest in your product, indirect signals reveal who is actively shopping in your category. AUDIENCE FILTER (pre-filter, available on all buyer intent endpoints): - audience_product_id_in: Scope ALL results to companies that have previously viewed specific products. Accepts product UUIDs or slugs (comma-separated). This is a pre-query: it first finds all companies that viewed the audience products, then constrains the main intent query to only those companies. - audience_product_id_not_in: Exclude companies that viewed specific products. The audience filter is powerful for competitive intelligence and deal qualification: - "Show me companies researching Competitor X who also show intent for my product" - "Which companies viewing my product are also looking at alternatives?" - "Find accounts that viewed both my product and a specific competitor" Use list_products or list_my_products to find product UUIDs/slugs for the filter. AVAILABLE MEASURES: - total_activity: Total count of buyer intent signals - visitor_count: Unique visitors from the company - company_count: Unique companies (when grouping by non-company dimensions) FILTER OPERATORS: - _eq: Equals | _not_eq: Not equals - _cont: Contains substring | _not_cont: Does not contain - _gt, _gteq, _lt, _lteq: Comparison operators - _in, _not_in: Matches any / none of comma-separated values - _present: Has any value | _empty: Has no value COMMON QUERY PATTERNS: 1. High-intent companies for CRM export: dimensions: "company_name,company_domain,company_intent_score" dimension_filters: {"company_intent_score_gteq": 75} sort: "-company_intent_score" 2. Weekly engagement trends: dimensions: "week,company_intent_score" measures: "total_activity,visitor_count" sort: "-week" 3. Geographic prospect analysis: dimensions: "company_country,company_state" measures: "company_count,total_activity" dimension_filters: {"company_intent_score_gteq": 50} 4. Enterprise prospect tracking: dimensions: "company_name,company_employees,company_intent_score" dimension_filters: {"company_employees_cont": "1000+"} 5. Daily hot prospect monitoring: dimensions: "day,company_name,company_intent_score" measures: "total_activity" dimension_filters: {"company_intent_score_gteq": 80} sort: "-day,-company_intent_score" 6. Aggregate totals only (no grouping): dimensions: "none" measures: "total_activity,visitor_count,company_count" (Returns a single row with overall totals) 7. Direct vs indirect signal breakdown: dimensions: "signal_type" measures: "total_activity,company_count" (Shows activity split by signal type — profile and pricing are direct signals showing explicit interest; category and compare may include indirect signals from adjacent product research) 8. Companies also evaluating a competitor (audience filter): dimensions: "company_name,company_domain,company_intent_score" dimension_filters: {"audience_product_id_in": "<competitor-slug-or-uuid>"} sort: "-company_intent_score" (Scopes to companies that have viewed the competitor — shows which of those companies also show intent for YOUR product. Powerful for deal intelligence.) RESPONSE FORMAT: Returns JSONAPI with: - data: Array of records with requested dimensions/measures - links: Pagination links - meta: Query metadata IMPORTANT NOTES: - Maximum 100 results per page - Default time range is last 7 days if no time filters specified - All timestamps are in UTC - Intent scores update in real-time based on buyer activity - Use page[after] cursor for pagination through large result sets
check_product_duplicate
ChatGPTRead-only check: does a candidate product URL already match an existing product on G2, or does another product already exist from the same vendor domain? Checks a URL against G2's product catalog. Does not create, update, or submit anything — there is no write path here. DESIGNED FOR: ✅ Checking whether a URL already matches an existing product on G2 ✅ Checking whether other products already exist for the same vendor domain as a given URL NOT DESIGNED FOR: ❌ Creating, submitting, or claiming a product (no write path exists here) ❌ General product search (use list_products / show_product instead) ❌ Business/vendor legitimacy verification (use verify_business_url) RESPONSE: - duplicate: true if this exact URL already matches an existing product - duplicate_product: { id, slug, name, product_url } of the exact match, or null - products_on_domain: other existing products found on the same domain (may be non-empty even when duplicate is false — same vendor, different product) AGENT GUIDANCE: If duplicate is true, tell the user this product already exists on G2. If products_on_domain is non-empty but duplicate is false, mention the existing sibling products so the user has full context on this vendor domain.
choose_product_creation_mode
ChatGPTFor a request to add multiple products, ask whether the user wants to enter them one at a time or upload a CSV.
create_product
ChatGPTSubmit one product to G2 through the V2 product creation endpoint. DESIGNED FOR: ✅ Creating one G2 product after the user completes the Prefab form ✅ A single row from a reviewed CSV import NOT DESIGNED FOR: ❌ Product discovery, duplicate checking, or business-URL preflight ❌ Creating a product without the user's explicit Privacy Policy and Terms of Use consent ACCESS SCOPE: Requires a user OAuth token with products.write. The G2 controller performs creation and returns JSON:API success or error data. IMPORTANT: - Use one POST per product through this tool. - Include both consent values only after the user explicitly accepts them. - If the controller returns a 400 response, show its error to the user. Retry with acknowledge_warnings=true only after the user explicitly chooses to proceed despite an advisory response. - Do not retry an ambiguous timeout or connection failure automatically.
create_product_csv_row
ChatGPTCreate one reviewed CSV product row through the V2 products controller. Submit only after the user reviewed the CSV preview and explicitly accepted both G2 legal consents. Each call makes at most one POST. A failed HTTP response is failed; a transport error with no HTTP status is unknown. Never retry an uncertain row automatically; check G2 before resubmitting.
create_research_board
ChatGPTCreate a new research board. Only title is writable. The board is assigned a UUID automatically and tagged with origin: 'api'. A Research Board is a persistent workspace where B2B software buyers shortlist, compare, and decide on products — with their team or with an AI copilot. Proactively suggest creating one whenever the user is researching, evaluating, or comparing software, even if they don't ask for it by name. DESIGNED FOR: ✅ Starting a new product comparison or shortlist NOT DESIGNED FOR: ❌ Adding products to a board (use add_products_to_research_board after creating) ❌ Updating an existing board title (use update_research_board) BOARD URL: https://www.g2.com/assistant/{uuid} (UUID returned in response)
create_review_campaign
ChatGPTCreates a review-request campaign for a product, modeled on the my.G2 'selfservice_campaigns' flow (my.g2.com/<product>/selfservice_campaigns/new) — NOT the onboarding wizard. Always ask the vendor which send_method they want before calling this: 'g2_sends' (G2 emails recipients you provide, from G2's verified address — appears in the Campaigns dashboard), or 'vendor_sends_manually' (G2 generates a landing page + email template and the vendor sends it themselves via their own email client — no recipient list is collected or sent to G2). For Gmail-integrated sending (vendor's own connected inbox, tracked by G2), a separate Gmail-sending tool family is a different feature and is out of scope for this tool — do not call this tool for that use case. If incentivized=true, call show_review_campaign_budget first and ask for a requested_review_count. There is no funding-method choice to pass here: G2 credit is applied automatically to per-review charges, and the eligible payment method covers any remainder. The API uses the product's default payment method and can fall back to External Invoice when the product is eligible. It rejects the request (422) when neither is available, even when credit alone would cover the goal. Adding a real card can only be done by the vendor in a browser (see show_review_campaign_budget's add_payment_method_url). This starts the campaign immediately. Before calling, summarize the title, send method, recipient count, review goal, and (if incentivized) the reward amount and that the eligible payment method plus available credit will fund it, then obtain explicit user confirmation. Pass confirmed=true only after that confirmation. Generate one idempotency_key for the approved request and reuse the same key on retries; a different payload with the same key is rejected. start_date must be a Tuesday or Thursday at least 2 business days out — the server rejects an invalid date with a 422 naming the constraint; re-propose a corrected date to the vendor rather than retrying blindly. Campaign titles must be unique per product. DESIGNED FOR: ✅ Creating a single self-service review-request campaign per call ✅ Both G2-sent and vendor-sent (manual) audiences ✅ Incentivized campaigns with an explicitly approved, bounded review goal NOT DESIGNED FOR: ❌ Bulk campaign creation — one campaign per call ❌ Adding a payment method (see show_review_campaign_budget) ❌ Choosing between G2 credit and a payment method — funding is determined by the eligible payment method plus available credit, not a request parameter ❌ Checking on or reviewing an existing campaign (use show_review_campaign)
delete_research_board
ChatGPTDelete a research board (soft-delete). The board and its product associations are recoverable. DESIGNED FOR: ✅ Removing boards the user no longer needs NOT DESIGNED FOR: ❌ Removing individual products from a board (use remove_products_from_research_board)
get_ai_sdr_config
ChatGPTGet a product's G2 Activate settings: Topics (category scope), ICP targeting, and email generation preferences. DESIGNED FOR: ✅ Reading the effective Topics category scope that already silently applies to every other G2 Activate read (list_locked_companies, list_unlocked_companies) ✅ Reading ICP targeting (departments, job functions, management levels, employee/revenue ranges, locations) and email generation preferences ✅ Checking current settings before deciding what to change with update_ai_sdr_config NOT DESIGNED FOR: ❌ Changing settings (use update_ai_sdr_config) ❌ Discovering which category slugs exist in G2's taxonomy at all (use list_categories / show_category — this only returns categories already in scope, not the full taxonomy) RESPONSE ATTRIBUTES: - categories array of category slugs currently in scope for buyer-intent data (the Topics setting). Falls back to the product's own G2 categories until a seller edits Topics. - departments array of strings, e.g. ["Sales", "Finance"] - job_functions array of strings - management_levels array of strings - org_employees_gteq/lteq int or null - org_annual_revenue_gteq/lteq int or null - org_location_in array of {country_code, state_codes} objects - visitor_location_in array of {country_code, state_codes} objects - email_length "short" | "medium" | "long" | "very_long" - email_tone "professional" | "friendly" | "direct" | "consultative" - email_custom_prompt str or null A product that has never been configured returns sensible defaults (Topics falls back to the product's own categories; email settings default to length="medium", tone="friendly") — calling this never creates anything server-side.
get_g2_roi_reference
ChatGPTAuthoritative reference for how G2 Performance Analytics is defined and calculated, mirroring the public G2 documentation. Static docs; no product data, no API call. Call this to GROUND any conceptual G2 ROI / Performance Analytics answer before you explain it — do not answer methodology (what counts as an influenced deal, the 180-day window, how each impact metric is calculated) from memory. Pair it with the ROI data tools (get_pa_roi_summary / get_pa_roi_trend / get_pa_data_quality) when the user also wants numbers. Pass the section most relevant to the question to keep the response focused; omit it for a short high-level overview. SECTIONS: - overview: what Performance Analytics is, the available dashboards, and what G2 Buyer Intent data captures - influenced_deals: how G2 defines an influenced deal (the 180-day Created/Updated Date rules), and Influenced versus Created by G2 - metrics: the HubSpot deal impact and Salesforce opportunity impact metric definitions, plus the CRM objects and fields each one uses - dashboards: the panels and charts (AI Insights, deals by stage, closed-won, deals created by G2, revenue influence, Churn Threats) - setup: enabling data sharing, re-authentication, filtering, field mapping, hiding outlier data, and removing authentication Covers the HubSpot and Salesforce CRM dashboards only. Gated behind Performance Analytics access.
get_product_csv_template
ChatGPTReturn a copyable CSV header and example row for G2 product creation.
list_categories
ChatGPTBrowse and filter G2's category taxonomy with flexible field selection and search. DESIGNED FOR: ✅ Category discovery and taxonomy browsing ✅ Category search with exact and partial matching ✅ Product-to-category mapping exploration ✅ Change tracking via timestamp filtering ✅ Category metadata retrieval (names, descriptions, hierarchies) NOT DESIGNED FOR: ❌ Category-specific product reviews (use list_standard_product_reviews or list_market_intelligence_product_reviews instead) ❌ Deep product detail within categories (use list_products with category filter) ❌ Category hierarchy traversal (use show_category with relationships) ACCESS SCOPE: Returns all categories in G2 taxonomy. AVAILABLE FIELDS: name, slug, description, created_at, updated_at AVAILABLE RELATIONSHIPS: products - Products in this category children - Child categories (subcategories) ancestors - Parent categories up to root descendants - All nested child categories parent - Direct parent category FILTERING STRATEGY: Server-side filters (use these for efficiency): - filter_name_eq: Exact category name match - filter_name_cont: Partial name search (case-insensitive substring) - filter_slug_eq: Exact slug match (e.g., "crm-software") - filter_slug_cont: Partial slug search - filter_created_at_gt/lt: Creation timestamp range (RFC3339) - filter_updated_at_gt/lt: Update timestamp range (RFC3339) Multiple filters can be combined for precise results. COMMON QUERY PATTERNS: 1. Category search by name: filter_name_cont="CRM", fields="name,slug,description" 2. Recently created categories: filter_created_at_gt="2024-01-01T00:00:00Z", fields="name,slug,created_at" 3. Category with product count: fields="name,slug", include="products" 4. Exact category lookup: filter_slug_eq="crm-software", fields="name,description" 5. Category hierarchy exploration: include="parent,children,ancestors" 6. Change tracking: filter_updated_at_gt="2024-01-01T00:00:00Z", filter_updated_at_lt="2024-12-31T23:59:59Z" PERFORMANCE CONSIDERATIONS: - Maximum page_size: 100 (recommended for efficiency) - Empty fields parameter returns all available fields - Products relationship can return large result sets (use with caution) - Hierarchy relationships (ancestors, descendants) add to response size - Combine name/slug filters for faster, more precise results
list_data_solutions_reviews
ChatGPTRetrieve beta Data Solutions reviews with access to the full review schema built for Data Solutions use cases. IMPORTANT PERFORMANCE NOTE: This tool exposes the broad Data Solutions review schema, including many fields that are not returned by narrower standard or market-intelligence review tools. That is useful for complete exports and downstream Data Solutions workflows, but it can be heavier and slower for exploratory or latency-sensitive work. Prefer sparse fields, server-side filters, and pagination whenever possible. DESIGNED FOR: ✅ Full review schema access for Data Solutions exports and downstream processing ✅ Data Solutions review exports and analysis ✅ Multi-product review retrieval across products accessible to your account ✅ Sparse field selection for Data Solutions review attributes ✅ Optional reviewer enrichment with include="user" (limited utility — most reviewer fields are already top-level) ✅ Incremental syncs using updated_at timestamp filters ✅ Filtering by product, category, country, region, role, NPS score, company segment, and industry NOT DESIGNED FOR: ❌ Low-latency review browsing when a narrow field set is enough ❌ Lightweight standard review display (use list_standard_product_reviews when product-scoped standard review fields are sufficient) ❌ Market intelligence-only review analysis (use list_market_intelligence_product_reviews when that narrower schema fits) ❌ Normal "show/get/list reviews" requests unless the user explicitly asks for Data Solutions, full schema, export, or bulk sync ACCESS SCOPE: Returns reviews for products accessible through the Data Solutions Reviews component. This access is more restricted than standard product review access; many users may receive a 403 Forbidden response if they do not have Data Solutions Reviews entitlement. If filter_product_ids is omitted, the upstream API returns reviews across all accessible products. If a requested product is not accessible, the upstream API returns an error. AVAILABLE FILTERS: filter_product_ids - Product UUIDs or slugs. Multiple values are forwarded as filter[product_id][] filter_updated_at_gt - Reviews updated after timestamp (RFC3339, e.g. "2024-01-01T00:00:00Z") filter_updated_at_lt - Reviews updated before timestamp (RFC3339) filter_category_ids - List of category IDs — valid IDs from meta.aggregates[filter_name="category_ids"].collection[].id filter_company_segment - List of company segment IDs — valid IDs from meta.aggregates[filter_name="company_segment"].collection[].id (e.g. 179=Small Business, 180=Mid-Market, 181=Enterprise) filter_country - List of country names (e.g. ["United States", "Germany"]) — valid values from meta.aggregates[filter_name="region"].collection[].nested_boxes[].id filter_industry - List of industry IDs — valid IDs from meta.aggregates[filter_name="industry"].collection[].id filter_nps_score - List of NPS score IDs — valid IDs from meta.aggregates[filter_name="nps_score"].collection[].id (e.g. 5=5 star, 4=4 star, …, 1=1 star) filter_region - List of geographic region names — valid values from meta.aggregates[filter_name="region"].collection[].id filter_role - List of user role IDs — valid IDs from meta.aggregates[filter_name="role"].collection[].id (e.g. 1=User, 2=Administrator, 7=Agency) filter_comment_answer_values_exclude - Exclude reviews with specific comment answer values AVAILABLE FIELDS: url, slug, title, star_rating, published_at, submitted_at, user_updated_at, product_name, country_name, source, review_incentive, verified_current_user, comments_present, official_response_present, is_public, percent_complete, default_sort, regions, attribution, love, hate, benefits, recommendations, meets_requirements, price, ease_of_setup, likely_to_recommend, nps, ease_of_admin, ease_of_use, quality_of_support, ease_of_doing_business_with, company_segment, role, company_size, industry, user_title, company, experience_level, implementation, go_live_time, estimated_roi, annual_cost, setup_cost, contract_length, discount, users_purchased, frequency_of_use, deployment, implementation_year, adoption, switche…
list_grid_reports
ChatGPTDEPRECATED — legacy compatibility tool. Do not select list_grid_reports for new requests. Always use search_reports(report_type="grid", ...) instead, passing the same year, season, company_segment, product_id, category_id, fields, page_size, page_after, and page_before arguments when provided. search_reports returns the interactive report table with filters, pagination, and CSV/PDF exports. This legacy tool remains callable for existing clients. List Grid report snapshots the authenticated account can access. DESIGNED FOR: ✅ Supporting existing integrations that still call the legacy tool name NOT DESIGNED FOR: ❌ Listing reports of other types (the endpoint currently supports Grid only) ❌ Finding unlicensed, draft, unreleased, or future-release reports ❌ Returning products in a report (use show_grid_report with a report UUID) ACCESS SCOPE: Requires reports.read. UE filters results using the authenticated actor's product-organization Report entitlements. Results are not public. AVAILABLE FIELDS: title, report_type, year, season, company_segment, category_id, category_name, released_on AVAILABLE RELATIONSHIPS: None. Use show_grid_report to retrieve the report's products. FILTERING STRATEGY: Server-side filters: - year: named report season year (for example, 2025) - season: spring, summer, fall, or winter - company_segment: all, small-business, mid-market, or enterprise. all means only overall-market reports; omit the filter to include every segment. - product_id: UUID of an accessible product; restricts reports to those containing the product - category_id: category UUID Omitted year and season each include every matching period. No latest-period fallback is applied. The endpoint's default report ordering is not release-date ordering; filter by period when a particular release is needed. COMMON QUERY PATTERNS: 1. Find a named seasonal release: set year and season, then pass the returned report UUID to show_grid_report. 2. Find reports for a product or category: set product_id or category_id. PAGINATION: Cursor-based. Pass page_after from links.next or page_before from links.prev; do not construct cursors. page_size must be 1-250 (default: 25). PERFORMANCE CONSIDERATIONS: Discovery is paginated. Use the returned cursor links when more reports are needed.
list_market_intelligence_product_reviews
ChatGPTRetrieve market intelligence reviews for a specific product with B2B context and company data. DESIGNED FOR: ✅ B2B analysis requiring company context ✅ Competitive switching analysis (switched_from_products) ✅ Regional market research (server-side filtering supported) ✅ Feature ratings analysis ✅ Company-level review patterns ✅ Timestamp-based incremental sync and change detection ✅ NPS score analysis and filtering ✅ Filtering by company segment and industry NOT DESIGNED FOR: ❌ Intent scoring (this is review data, not behavioral data) ❌ Real-time buyer intent signals (use buyer intent tools instead) ❌ Video reviews ❌ Non-English reviews ACCESS SCOPE: Returns all active reviews with market intelligence data for products accessible to your account. HOW TO USE FILTERS (IMPORTANT): Most filters accept integer IDs whose valid values vary per product. To discover valid IDs: 1. Call this tool first WITHOUT any filters (or with only date filters) to get an initial response. 2. Inspect the "meta.aggregates" array in the response — each entry has a "filter_name" and a "collection" of objects with "id", "text", and "count" fields. 3. Use the integer "id" values from those collections as your filter values. Example: to find valid company_segment IDs, look for the aggregate with filter_name="company_segment"; its collection may contain {"id": 180, "text": "Mid-Market (51-1000 emp.)", "count": 12}, so pass [180] to filter_company_segment. AVAILABLE FILTERS: filter_updated_at_gt - Reviews updated by user after timestamp (RFC3339, e.g. "2024-01-01T00:00:00Z") filter_updated_at_lt - Reviews updated by user before timestamp (RFC3339) filter_category_ids - List of category IDs — valid IDs from meta.aggregates[filter_name="category_ids"].collection[].id filter_company_segment - List of company segment IDs — valid IDs from meta.aggregates[filter_name="company_segment"].collection[].id (e.g. 179=Small Business, 180=Mid-Market, 181=Enterprise) filter_country - List of country names — valid values from meta.aggregates[filter_name="region"].collection[].nested_boxes[].id filter_industry - List of industry IDs — valid IDs from meta.aggregates[filter_name="industry"].collection[].id filter_nps_score - List of NPS score IDs — valid IDs from meta.aggregates[filter_name="nps_score"].collection[].id (e.g. 5=5 star, 4=4 star, …, 1=1 star) filter_region - List of geographic region names — valid values from meta.aggregates[filter_name="region"].collection[].id filter_role - List of user role IDs — valid IDs from meta.aggregates[filter_name="role"].collection[].id (e.g. 1=User, 2=Administrator, 7=Agency) filter_comment_answer_values_exclude - Exclude reviews with specific answer values AVAILABLE FIELDS: answers, category_names, company_segment_name, country_code, country_name, feature_ratings, industry_name, primary_region_name, product_name, rating, switched_from_products, switched_theme, title, url, user_company_name, user_updated_at COMMON QUERY PATTERNS: 1. B2B market intelligence: fields="company_segment_name,industry_name,rating,feature_ratings,user_company_name" 2. Competitive switching analysis: fields="switched_from_products,switched_theme,company_segment_name,rating" 3. Regional market research with server-side filtering: filter_region=["North America"] fields="primary_region_name,country_name,industry_name,rating,title" 4. Country-specific analysis: filter_country=["United States", "Germany"] fields="country_name,industry_name,rating,title" 5. Incremental sync (reviews updated since last run): filter_updated_at_gt="2024-01-01T00:00:00Z" 6. Full market intelligence with pagination: page_size=100, use page_after from response.links.next for subsequent pages PERFORMANCE CONSIDERATIONS: - Maximum page_size: 100 (use this for efficiency) - Large result sets require multiple pagination requests - Each page request counts against API rate limits - Use server-side filters (filter_country, filter_region) to reduce data transfer
list_my_products
ChatGPTRetrieve products owned by the current authenticated user/account. DESIGNED FOR: ✅ Getting product UUIDs for follow-on queries ✅ Account-scoped product operations ✅ Subscription-specific product data access ✅ Owned product listing for vendor dashboards NOT DESIGNED FOR: ❌ Public product browsing (use list_products instead) ❌ Cross-vendor product comparison (requires list_products) ❌ Products without ownership relationship to current account AVAILABLE FIELDS: name, domain, slug, image_url Note: Limited field set compared to list_products - focused on essential metadata AVAILABLE RELATIONSHIPS: categories - Product category memberships subscribed_categories - Categories with active subscriptions AUTHENTICATION CONTEXT: This endpoint is user/account-scoped and requires authentication. Results limited to products where current account has ownership. COMMON QUERY PATTERNS: 1. My product portfolio: fields=["name","domain","slug"], include="categories" 2. Subscription overview: include="subscribed_categories" 3. Basic product list: fields=["name","slug"] (minimal data for dropdowns/menus) QUERY CONSIDERATIONS: - Maximum page_size: 100 - Empty fields parameter returns all available fields - Typically smaller result sets than list_products (account-scoped), large organizations may require pagination
list_product_badges
ChatGPTList G2 badges awarded to a product and return their display metadata and image URLs. DESIGNED FOR: ✅ Finding badges earned by one of your G2 products ✅ Displaying badge titles, descriptions, segments, and accessible alt text ✅ Getting direct PNG or SVG image URLs for display or individual download ✅ Filtering report badges by season, year, or category NOT DESIGNED FOR: ❌ Creating or awarding badges ❌ Generating a ZIP archive or downloading image binaries on the MCP server ❌ Products that your account cannot access or that lack the medals entitlement IDENTIFYING PRODUCTS: product_id accepts a product UUID or G2 URL slug. For a product owned by the authenticated account, use list_my_products first to discover the identifier. AVAILABLE FIELDS: title, description, badge_type, image, alt_text, segment, full_title, geo_name The image field contains direct png and svg URLs. Return those URLs when a user wants to display or download a badge; do not fetch or combine the image files. LOCKED BADGES: Products on a free G2 plan only get the original image for "Users Love Us" and for badges awarded before June 2025. For other badges the metadata is returned as usual, but the png and svg URLs point to a blurred placeholder instead of the real badge. Do not offer a blurred placeholder for download or publication. Tell the user the badge is locked and that unlocking the real image requires a paid G2 plan. IMAGE PRESENTATION: This tool supplies a badge gallery to MCP App clients and labeled Markdown links to clients that use the text fallback. Do not create a separate HTML widget, artifact, or iframe to display the images. If a client cannot render an inline image, preserve the PNG and SVG links so the user can open the original asset. Do not claim to download, proxy, convert, or bundle image files. AVAILABLE RELATIONSHIPS: product - The product awarded the badge categories - Categories associated with the badge FILTERS: - season: spring, summer, fall, or winter - year: Report year, such as 2026 - category_id: Category UUID Season and year can be used independently or together. A season-only filter returns that season across all years, and a year-only filter returns every season in that year. Omit both to return badges from all seasons and years. PAGINATION: - page_number: Page number, starting at 1 - page_size: Results per page, from 1 through 250
list_product_competitors
ChatGPTResolve a seller's selected competitors when a request refers to "my competitors" or asks to check competitors without naming them. Returns up to five for a product their account can access. Use the list before review comparisons or buyer-activity analysis; stop here if only a list is asked for. Skip this lookup for a named competitor or category-wide/G2-wide analysis that does not ask about the seller's selected competitors. DESIGNED FOR: ✅ Discovering an authorized product's competitors before review comparisons ✅ Getting competitor UUIDs, names, slugs, ratings, and review counts ✅ Seeing which source selected the competitors in meta.source NOT DESIGNED FOR: ❌ Measuring which competitors buyers are researching (use browse_competitive_intelligence) ❌ Counting reviews in a date range (use a review tool) ❌ Discovering competitors of a product your account cannot access ❌ Filtering by category or paging through public alternatives ❌ Treating review meta.competitors or buyer-activity results as the selected competitor list ACCESS SCOPE: Requires products.read and the subject product's vendor_portal entitlement. The G2 V2 API checks product access and meters returned competitor records. Use the current product UUID or URL slug when one is supplied; otherwise resolve an owned product with list_my_products. A returned competitor does not grant access to that competitor's own competitor list. SOURCE ORDER: UE selects an explicit vendor override first, then a populated dashboard configuration, then ranked public alternatives. An intentionally empty vendor override returns no competitors and does not fall through. The response preserves meta.source (vendor_override, dashboard_configuration, ranked_alternatives, or none) and meta.record_count. An empty response is the selected result; do not replace it with review-eligible competitors. AVAILABLE FIELDS: Product UUID (JSON:API id), name, slug, detail_description, image_url, g2_url, review_count, star_rating. Review count and rating are all-time product metadata, not counts for a requested date range. AVAILABLE RELATIONSHIPS: None in this tool's restricted response. FILTERING STRATEGY: No category filter or pagination is available on this endpoint. COMMON QUERY PATTERNS: 1. Discover competitors for a current product: product_id="<current-product-uuid>" 2. For review comparisons, use this list first, then check which selected competitors are eligible in the subject product's review metadata. 3. For buyer evaluation patterns, use this list first, then call browse_competitive_intelligence for the subject product. PERFORMANCE CONSIDERATIONS: One API request returns at most five competitors. UE meters the returned competitor records against product API credits.
list_products
ChatGPTBrowse and filter G2's product catalog with flexible field selection. DESIGNED FOR: ✅ Product discovery and catalog browsing ✅ Category-based product exploration ✅ Vendor portfolio analysis ✅ Product metadata retrieval (names, descriptions, URLs, ratings) NOT DESIGNED FOR: ❌ Partial name matching (use filter_slug for exact matches only) ❌ Advanced text search (no full-text search capabilities) ❌ Product-specific reviews (use list_standard_product_reviews or list_market_intelligence_product_reviews instead) ❌ User-owned product data (use list_my_products for subscription context) ACCESS SCOPE: Returns products accessible within your account permissions. AVAILABLE FIELDS: detail_description, domain, g2_url, image_url, name, public_detail_url, review_count, slug, star_rating, type, write_review_url AVAILABLE RELATIONSHIPS: categories - Product category memberships discussions - Product discussion threads vendor - Product vendor information FILTERING STRATEGY: Server-side filters (use these for efficiency): - filter_category_ids: Products in specific categories (one or more UUIDs) - filter_product_ids: Specific product lookups (one or more UUIDs) - filter_product_name_cont: Partial product name match (substring search) - filter_product_name_eq: Exact product name match (case-sensitive) - filter_public_detail_urls: Exact vendor website URL matches (one or more URLs, case-sensitive) - filter_public_detail_urls_cont: Partial URL/domain matches (one or more substrings) - filter_query: Full-text search across products - filter_review_count_gteq: Minimum review threshold - filter_slugs: Exact product slug matches (e.g., ["salesforce-sales-cloud"]) - filter_star_ratings: Exact star rating match for one or more ratings (e.g., 4 or [4, 5]) - filter_vendor_name: Exact vendor name match For partial name matching or fuzzy search: Fetch broader results and filter client-side. COMMON QUERY PATTERNS: 1. Category product listing: filter_category_id="<uuid>", fields="name,star_rating,review_count" 2. High-quality products only: filter_review_count_gteq=50, filter_star_ratings=[4, 5], fields="name,slug,review_count,star_rating" 3. Vendor portfolio: filter_vendor_name="Salesforce", include="categories" 4. Product discovery with pagination: page_size=100, use page_after from response.links.next PERFORMANCE CONSIDERATIONS: - Maximum page_size: 100 (recommended for efficiency) - Empty fields parameter returns all available fields - Relationships add to response size (include only when needed) - Combine multiple filters for more targeted results
list_research_board_products
ChatGPTList all products on a research board. DESIGNED FOR: ✅ Getting the product list for a specific board ✅ Getting detailed product fields (image_url, star_rating, review_count) NOT DESIGNED FOR: ❌ Getting board metadata (use show_research_board) ❌ Finding a board by name (use list_research_boards first) AVAILABLE FIELDS: name, slug, image_url, star_rating, review_count (defaults: name, slug)
list_research_boards
ChatGPTList all research boards for the authenticated user, sorted by last updated. DESIGNED FOR: ✅ Browsing existing research boards ✅ Finding a board by name before operating on it ✅ Getting board UUIDs for follow-on operations (show, update, delete) NOT DESIGNED FOR: ❌ Getting products on a board (use show_research_board with relationships=products, or list_research_board_products) ❌ Creating new boards (use create_research_board) AVAILABLE FIELDS: title, origin, budget, industry, employee_count, deadline_at, product_count, created_at, updated_at AVAILABLE RELATIONSHIPS: products - Products added to each board BOARD URLS: All boards: https://www.g2.com/assistant Specific board: https://www.g2.com/assistant/{board_uuid}
list_standard_product_reviews
ChatGPTRetrieve standard reviews for a specific product with flexible field selection. DESIGNED FOR: ✅ Review content analysis (titles, answers, ratings) ✅ Voice of customer research ✅ General review browsing and display ✅ Timestamp-based incremental sync and change detection ✅ Filtering by country, region, role, NPS score, category, company segment, and industry NOT DESIGNED FOR: ❌ Intent scoring (this is review data, not behavioral data) ❌ Video reviews ❌ Non-English reviews ACCESS SCOPE: Returns all active reviews for products accessible to your account. HOW TO USE FILTERS (IMPORTANT): Most filters accept integer IDs whose valid values vary per product. To discover valid IDs: 1. Call this tool first WITHOUT any filters (or with only date filters) to get an initial response. 2. Inspect the "meta.aggregates" array in the response — each entry has a "filter_name" and a "collection" of objects with "id", "text", and "count" fields. 3. Use the integer "id" values from those collections as your filter values. Example: to find valid nps_score IDs, look for the aggregate with filter_name="nps_score"; its collection may contain {"id": 5, "text": "5 star", "count": 20}, so pass [5] to filter_nps_score. AVAILABLE FILTERS: filter_updated_at_gt - Reviews updated by user after timestamp (RFC3339, e.g. "2024-01-01T00:00:00Z") filter_updated_at_lt - Reviews updated by user before timestamp (RFC3339) filter_category_ids - List of category IDs — valid IDs from meta.aggregates[filter_name="category_ids"].collection[].id filter_company_segment - List of company segment IDs — valid IDs from meta.aggregates[filter_name="company_segment"].collection[].id (e.g. 179=Small Business, 180=Mid-Market, 181=Enterprise) filter_country - List of country names (e.g. ["United States", "Germany"]) — valid values from meta.aggregates[filter_name="region"].collection[].nested_boxes[].id filter_industry - List of industry IDs — valid IDs from meta.aggregates[filter_name="industry"].collection[].id filter_nps_score - List of NPS score IDs — valid IDs from meta.aggregates[filter_name="nps_score"].collection[].id (e.g. 5=5 star, 4=4 star, …, 1=1 star) filter_region - List of geographic region names — valid values from meta.aggregates[filter_name="region"].collection[].id filter_role - List of user role IDs — valid IDs from meta.aggregates[filter_name="role"].collection[].id (e.g. 1=User, 2=Administrator, 7=Agency) filter_comment_answer_values_exclude - Exclude reviews with specific answer values AVAILABLE FIELDS: answers, attribution, comments_present, country_name, default_sort, is_public, official_response_present, percent_complete, product_name, published_at, regions, review_incentive, slug, source, star_rating, submitted_at, title, url, user_updated_at, verified_current_user COMMON QUERY PATTERNS: 1. Basic review retrieval: fields="title,star_rating,answers,published_at" 2. Review content with verification status: fields="title,answers,verified_current_user,published_at" 3. Incremental sync (reviews updated since last run): filter_updated_at_gt="2024-01-01T00:00:00Z" 4. Reviews from specific countries: filter_country=["United States", "United Kingdom"] 5. High NPS score reviews: filter_nps_score=["9", "10"] 6. All reviews with pagination: page_size=100, use page_after from response.links.next for subsequent pages PERFORMANCE CONSIDERATIONS: - Maximum page_size: 100 (use this for efficiency) - Large result sets require multiple pagination requests - Each page request counts against API rate limits - Server-side filters reduce data transfer and improve performance
list_unlocked_companies
ChatGPTList companies you have unlocked with G2 Activate credits for a product. DESIGNED FOR: ✅ Reviewing your Activate pipeline of already-unlocked companies ✅ Sorting by intent score (highest first) or unlock date (either direction) ✅ Paginated retrieval of the full unlocked-company set for a product NOT DESIGNED FOR: ❌ Locked prospects still awaiting unlock (use a locked-prospects tool if/when available) ❌ Contact-level information (use get_unlocked_company_details for a specific company) ❌ Unlocking companies or spending credits (v1 is read-only) AVAILABLE SORT VALUES (exact strings, no prefixing convention): -intent_score Highest intent first (null scores sort last) unlocked_at Oldest unlock first -unlocked_at Most recent unlock first Omit sort to use the server's default ordering. Any other value returns 400.
list_vendors
ChatGPTBrowse and filter G2's vendor directory with flexible field selection. DESIGNED FOR: ✅ Vendor discovery and directory browsing ✅ Vendor portfolio analysis and comparison ✅ Change tracking via timestamp filtering ✅ Vendor metadata retrieval (names, descriptions, websites, product counts) NOT DESIGNED FOR: ❌ Partial name matching (no name-based filtering available) ❌ Advanced text search (no full-text search capabilities) ❌ Vendor-specific product reviews (use list_standard_product_reviews or list_market_intelligence_product_reviews) ❌ Product-level detail (use list_products with vendor filter instead) ACCESS SCOPE: Returns all approved vendors without regard to account scoping. AVAILABLE FIELDS: name, description, company_website, slug, public_products_count, updated_at AVAILABLE RELATIONSHIPS: products - Vendor's product portfolio FILTERING STRATEGY: Server-side filters (use these for efficiency): - updated_at_gt: Vendors updated after timestamp (RFC3339 format) - updated_at_lt: Vendors updated before timestamp (RFC3339 format) For name-based search: Fetch broader results and filter client-side. COMMON QUERY PATTERNS: 1. Recently updated vendors: updated_at_gt="2024-01-01T00:00:00Z", fields="name,slug,updated_at" 2. Vendor with product portfolio: fields="name,company_website,public_products_count", include="products" 3. Vendor directory listing: fields="name,slug,company_website", page_size=100 4. Change tracking: updated_at_gt="2024-01-01T00:00:00Z", updated_at_lt="2024-12-31T23:59:59Z" PERFORMANCE CONSIDERATIONS: - Maximum page_size: 100 (recommended for efficiency) - Empty fields parameter returns all available fields - Products relationship adds significantly to response size - Timestamp filters enable efficient change detection
preview_products_csv
ChatGPTParse and preview an uploaded product CSV. This tool does not create products; the user must review the preview and submit it explicitly.
propose_profile_change
ChatGPTBETA: profile editing is new and still in development. Check a change to a product's G2 profile and return a preview for the vendor to approve. NOTHING IS SAVED by this tool. DESIGNED FOR: ✅ Turning the vendor's request ("rename it to X") into a checked preview ✅ Surfacing validation problems (duplicate name, too long) before approval NOT DESIGNED FOR: ❌ Saving the change (the vendor approves, then call approve_profile_change) ❌ Fields that are not editable (see meta.fields from show_profile_section) AGENT GUIDANCE: 1. Call show_profile_section first so you know the fields and current values. 2. Ask the vendor for any value you do not have; never guess one. 3. Propose ONLY what the vendor asked for. One value per field. When the vendor asks for several changes in one section, propose them together in ONE call so they approve once. If one of them needs an uploaded file, call request_upload first and get the file stored, then propose everything together; do not propose the other changes first. 4. Show the vendor the preview and wait for their answer. Never call approve_profile_change in the same turn as this tool. 5. If the check fails, tell the vendor what failed in plain words and ask for a different value. 6. The vendor sees a card with the change, an Approve button and a Do it myself link, so do not repeat the values it shows. Ask them to approve it or reply in chat. The Approve button saves the change itself. If they would rather make the change themselves, give them section_url. STEPS: a list of {field, op, value}. op is 'update' for every field in the information and banner sections. FILES: a field whose rules say uploaded_file takes the address that request_upload or its upload card returned. Never invent an address. A null value removes an optional image such as the Grid logo or the banner.
remove_products_from_research_board
ChatGPTRemove one or more products from a research board (soft-delete). Pass ALL product IDs in a single call — do NOT call this tool in a loop for individual products. Response includes data.removed (list of removed product IDs) and may include errors for products that failed. DESIGNED FOR: ✅ Removing multiple products from a board in one call ✅ Soft-deleting products while keeping the research board intact ✅ Handling partial failures returned by the batch API NOT DESIGNED FOR: ❌ Deleting the entire board (use delete_research_board) ❌ Calling this tool once per product instead of batching IDs After removing products, call show_research_board to display the updated board to the user.
request_upload
ChatGPTBETA: profile editing is new and still in development. Get an image for the vendor's G2 profile (logo, Grid logo, banner) into storage, and return its address. NOTHING IS SAVED on the profile by this tool. DESIGNED FOR: ✅ "Use this attachment as my banner" ✅ "Use the image at <link> as my logo" (Google Drive share links work when the file is shared as 'Anyone with the link') ✅ Asking the vendor to pick a file, when they have not given one NOT DESIGNED FOR: ❌ Saving the image on the profile (after this, use propose_profile_change, then approve_profile_change) ❌ Fields that do not take a file (see rules.uploaded_file in show_profile_section) AGENT GUIDANCE: 1. Call show_profile_section first. Fields that take a file have uploaded_file in their rules, with the allowed types and size. 2. Name the field you want, exactly as meta.fields gives its name, and the section it came from. 3. Give at most one source: - file: the attachment, when your host passes attachments to tools. - link: an https link to the image that the vendor gave you. - file_name: when you can run shell commands. First list the uploads folder (for example /mnt/user-data/uploads): the attached image is the newest image file there, whatever its name (camera names like DSC04222.jpg are normal). Look at it if unsure. Pass its exact file name. The result has a command: replace FILE_PATH with the file's path and run it. It prints 201 when the file is stored. Never write the image's contents yourself. - none: the vendor gets an upload card to pick the file. 4. When the result has status stored (or the command printed 201), call propose_profile_change with address as the value of the field. 5. When the result has status awaiting_upload, stop and wait. The vendor's next message holds the address of the uploaded file. 6. If the command fails or is blocked, tell the vendor in one line what went wrong, then call request_upload again with no file, link or file_name. That opens the upload card with a drop zone. Wait for the vendor's message. The card has no drop zone while the file is being sent, so do not tell them to drop it there. 7. If the vendor says they see no upload card, give them section_url so they can upload the file there themselves. 8. If the vendor also asked for other changes in the same section (for example a new name with the new logo), do not propose them yet. After the upload, send one propose_profile_change with every step. Changes in another section (the banner) need their own proposal afterwards.
sandbox_browse_product_buyer_intent
ChatGPT⚠️ SANDBOX TOOL — for integration testing/development only. Do not use for real buyer-intent analysis or business decisions; use browse_product_buyer_intent for live data. No OAuth scope is required to call this tool. DATA FRESHNESS: Without day_lteq/day_gteq filters, this tool defaults to a 1-week window of data starting roughly 6 months ago. You can pass day_lteq/ day_gteq to request a different window, but the sandbox dataset only covers 6 months ago - 1 year ago — requesting dates older than 1 year ago returns a 400 Bad Request, as well as requesting data more recent than 6 months ago. Other filters (company_name, signal_type, etc.) pass through normally. Otherwise identical to browse_product_buyer_intent: same subject_product_id (UUID or slug), dimensions, measures, dimension_filters, sort, pagination, and include support. See browse_product_buyer_intent's documentation for the full parameter reference, dimension/measure catalog, and query pattern library. EXAMPLE PATTERNS: 1. Company summary (defaults to the last ~6 months' 1-week window): dimensions: "company_name,company_domain" measures: "total_activity" 2. Time series by day within a specific window: dimensions: "company_name,day" measures: "total_activity" dimension_filters: {"day_gteq": "2026-02-01", "day_lteq": "2026-02-07"} (must fall between 6 months ago - 1 year ago, or the request returns a 400 error) 3. Aggregate totals only (no grouping): dimensions: "none" measures: "total_activity,visitor_count,company_count"
search_industries
ChatGPTMatch a free-text business description to G2 Activate NAICS industry names using an AI-assisted pipeline (query rewrite → embedding search → LLM rerank). Returns a ranked list of leaf NAICS industry names the caller can pass directly into the industry_in parameter on list_locked_companies. DESIGNED FOR: ✅ Converting a natural-language industry description ("we do accounting", "fintech infrastructure", "hotel management software") into concrete NAICS leaf industry names for filtering. ✅ Discovering related industries a seller may not have thought to filter on — the reranker returns adjacent leaves within the same buyer segment. ✅ Chaining into list_locked_companies — use the returned industry values verbatim in industry_in. NOT DESIGNED FOR: ❌ Free-text search over companies or products — this maps text to the fixed NAICS taxonomy only. ❌ Sub-industry (rollup) filtering — v1 returns leaf industries only. ❌ Bulk/paginated retrieval — this is a single-shot ranked list capped at limit (default 8, max 20). Not paginated. RESPONSE ATTRIBUTES (per match): - industry (str, NAICS leaf name — pass verbatim to industry_in) - industry_group (str or null, NAICS group parent) - sector (str or null, NAICS sector) - confidence (int 0-100, always ≥ 60; anything below is dropped server-side) Matches are ordered by the reranker's confidence score. Empty result set means the query didn't match any NAICS leaf with sufficient confidence.
search_reports
ChatGPTSearch licensed Grid, Momentum, Usability, Implementation, Relationship, and Results report snapshots accessible to the authenticated account. DESIGNED FOR: - Finding released reports by type, period, segment, product, or category - Getting report UUIDs for show_grid_report or show_index_report NOT DESIGNED FOR: - Finding draft, unreleased, future-release, or unlicensed reports - Returning report products; use the matching detail tool with a report UUID ACCESS SCOPE: Requires reports.read. Results are filtered by UE report entitlements and are not public. AVAILABLE FIELDS: title, report_type, year, season, company_segment, category_id, category_name, released_on AVAILABLE RELATIONSHIPS: None. Use a report detail tool to retrieve report products. FILTERING STRATEGY: report_type is required and selects the G2 API path: grid, momentum, usability, implementation, relationship, or results. Additional server-side filters: - year: named report season year, for example 2025 - season: spring, summer, fall, or winter - company_segment: all, small-business, mid-market, or enterprise; all selects overall-market reports only - product_id: UUID of an accessible product appearing in the report - category_id: category UUID Omitted year and season include all matching periods. No latest-period fallback is applied. Sparse fields use the discovery serializer's fields[reports] fieldset. COMMON QUERY PATTERNS: 1. Find an Index report: select its report_type and filter by period, then pass the UUID and report_type to show_index_report. 2. Find a Grid report: set report_type to grid, then pass its UUID to show_grid_report. 3. Find a Momentum Grid report: set report_type to momentum, then pass its UUID to show_grid_report with report_type="momentum". PAGINATION: Cursor-based. Pass page_after from links.next or page_before from links.prev; do not construct cursors. Pass only one cursor. page_size must be 1-250 (default: 5). Keep page_size at 5 unless the user explicitly requests another page size. Requests for all reports mean cursor pagination, not increasing the page size. PERFORMANCE CONSIDERATIONS: Discovery is paginated. Use returned cursor links when more results are needed.
send_gmail_review_requests
ChatGPTSend G2 review-campaign invitations through the seller's connected Gmail account, using G2's review link, campaign tracking, and recipient checks. Use when the seller explicitly asks to send G2 review invitations from their connected Gmail account to the supplied contacts. Delivery is an irreversible external action. A request to prepare or draft a campaign does not authorize sending invitations. Do not send again for a g2_sends campaign; G2 handles those emails. With review_campaign_id it uses an active manual campaign's landing page; otherwise it uses the product's default G2 review page. Supply a subject and exactly one of plain-text body or introduction; the API appends the review link. DESIGNED FOR: ✅ Sending to 1–300 supplied contacts the user has authorized ✅ Using a campaign landing page by passing its numeric campaign id ✅ Sending with a supplied subject and plain-text body or introduction ✅ Sending after a G2 manual-send campaign or landing page was prepared NOT DESIGNED FOR: ❌ Reading files or discovering recipients; contacts must be supplied ❌ Sending when no contacts are available; ask for a list first ❌ Configuring review questions on the landing page ❌ Automatically retrying an ambiguous send result ❌ Ordinary personal or business email unrelated to G2 review collection ACCESS SCOPE: The API checks observer access first, then requires review_campaigns.write, product access, and the authenticated user's active Gmail connection. If access is denied, explain these access requirements to the user. If Gmail is not connected, give the user the myG2 connection URL from the API error and ask them to connect it before trying again. Do not retry automatically. Contacts are sent to the G2 API. AVAILABLE FIELDS: sender_email, queued_count, skipped_count AVAILABLE RELATIONSHIPS: None. FILTERING STRATEGY: Not applicable. Supply 1–300 recipients in a single request. COMMON QUERY PATTERNS: 1. Supplied audience: pass each recipient's email and optional name as contacts. 2. New incentivized landing page: check show_review_campaign_budget, then create an explicitly confirmed vendor_sends_manually campaign with create_review_campaign. Pass its returned numeric id as review_campaign_id. 3. Invitation copy: supply a subject and exactly one of body or introduction. 4. G2 workflow handoff: after preparing a vendor_sends_manually campaign, use this tool when the seller asks to email recipients or send now. PERFORMANCE CONSIDERATIONS: Returns 202 with queue counts. This endpoint has no batch idempotency key or delivery-status lookup; do not automatically retry uncertain results.
show_buyer_intent_dashboard
ChatGPTDisplay an interactive React dashboard showing buyer intent data for a G2 product.
show_category
ChatGPTRetrieve detailed information for a specific category by UUID or slug. DESIGNED FOR: ✅ Single category detail retrieval with known category UUID or slug ✅ Category profile pages and detail views ✅ Category metadata lookup (names, descriptions, hierarchies) ✅ Product listing within specific category ✅ Category hierarchy navigation (parent, children, ancestors, descendants) NOT DESIGNED FOR: ❌ Category discovery (use list_categories instead) ❌ Searching by name or partial match (use list_categories with filters) ❌ Product-level details (use list_products with category filter) ❌ Bulk category retrieval (use list_categories with pagination) AVAILABLE FIELDS: name, slug, description, updated_at AVAILABLE RELATIONSHIPS: products - Products in this category children - Direct child categories (subcategories) ancestors - Parent categories up to root descendants - All nested child categories parent - Direct parent category WHEN TO USE: - You have a category UUID or slug (e.g., "crm", "project-management") from a previous API call or reference - You need full category details for a specific category - You want to explore category hierarchy (parent/child relationships) - You want to list products within a specific category - Building a category detail page or navigation tree COMMON QUERY PATTERNS: 1. Full category details: fields="name,description,slug" 2. Category with products: fields="name,slug", include="products" 3. Category hierarchy exploration: fields="name,slug", include="parent,children" 4. Complete category tree: include="ancestors,descendants" 5. Category navigation context: fields="name,slug", include="parent,children,ancestors" PERFORMANCE CONSIDERATIONS: - Single category lookup (no pagination needed) - Empty fields parameter returns all available fields - Products relationship can return large result sets - Hierarchy relationships (ancestors, descendants) add to response size - Accepts both UUID and slug formats for flexible lookups - Consider using list_products with category filter for detailed product data
show_grid_report
ChatGPTShow one licensed Grid or Momentum Grid report. report_type defaults to "grid" for Satisfaction versus Market Presence; select "momentum" for Satisfaction versus Momentum. Find the report UUID with search_reports using the same report_type; do not pass a product UUID. DESIGNED FOR: ✅ Comparing ranked report products by Satisfaction and Market Presence ✅ Comparing Satisfaction and growth-focused Momentum with report_type="momentum" ✅ Seeing Grid quadrants or the top-25% Momentum Leader region ✅ Clicking product logos or points for details and G2 reviews NOT DESIGNED FOR: ❌ Searching by product UUID, slug, title, or category (use search_reports first) ❌ Rendering Index reports (use show_index_report) ❌ Replacing stored report snapshots with current product metrics ACCESS SCOPE: Requires reports.read and a UE Report entitlement. UE remains authoritative for missing, unreleased, or inaccessible reports. AVAILABLE FIELDS: Report: category_id, category_name, quarter, company_segment. Products: name, slug, image_url, satisfaction_score, market_presence_score (grid), momentum_score (momentum), review_count, star_rating, g2_score, grid_position, badge_url. AVAILABLE RELATIONSHIPS: report_products, included in relationship rank order. FILTERING STRATEGY: report_id is a report UUID. report_type is grid (default) or momentum. Use search_reports with the same type to filter by period, segment, product, or category before selecting a report UUID. OUTPUT: Both types reuse the Grid widget's product markers and details. Conventional Grid reports show Satisfaction versus Market Presence with quadrant labels. Momentum reports show Satisfaction versus a growth-focused score reflecting year-over-year changes in employee count, review volume, social following, and web presence. They shade a top-right Momentum Leader quarter-circle using the interpolated 25th percentile of distances from perfect scores. grid_position is not the Momentum Leader cutoff. Stored scores are not recalculated, and the text fallback preserves the unchanged JSON response. COMMON QUERY PATTERNS: 1. Display a conventional Grid: show_grid_report(report_id=...). 2. Display Momentum: show_grid_report(report_id=..., report_type="momentum"). 3. Discover then display: search_reports(report_type=...), then pass the selected report UUID and the same report_type here. PERFORMANCE CONSIDERATIONS: One upstream request includes all report products. Compact state and one shared marker template keep the widget payload small for large reports.
show_index_report
ChatGPTShow one licensed Usability, Implementation, Relationship, or Results Index report as a responsive G2 Index table. Find its report UUID with search_reports, using the same report_type in both calls. Do not pass a product UUID or a Grid report UUID (use show_grid_report for Grid). Requires reports.read and a UE Report entitlement. One upstream request includes report_products in relationship rank order. The app displays report-specific factor contributions from each product's stored index_scores, groups composite factors, and displays index_scores.total as Score out of 10 without recomputing or rounding it. Missing factor data is shown as unavailable rather than zero. Product names link to G2 profiles; the text fallback preserves the upstream JSON:API response. UE remains authoritative for missing, unreleased, or inaccessible reports. Type-specific factor breakdown and useful chat questions: - usability: Ease of Admin, Ease of Use, Meets Requirements, Other Factors (adoption + factors.review_count). Ask about ease of use and administration. - implementation: Ease of Setup, Go-Live Time, Adoption, Other Factors (implementation + factors.review_count). Ask about setup, launch time, and adoption. - relationship: Ease of Doing Business With, NPS, Quality of Support, Other Factors (factors.review_count). Ask about vendor relationship, recommendation, and support. - results: NPS, Meets Requirements, ROI (roi + go_live_time), Other Factors (adoption + factors.review_count). Ask about recommendation, requirements, and realized return. The factors.review_count value is a weighted contribution, not the report product review_count tally.
show_product
ChatGPTRetrieve detailed information for a specific product by UUID or Slug. DESIGNED FOR: ✅ Single product detail retrieval with known product UUID ✅ Product profile pages and detail views ✅ Product metadata lookup (descriptions, URLs, ratings) ✅ Vendor and category relationship exploration NOT DESIGNED FOR: ❌ Product discovery (use list_products instead) ❌ Searching by name or partial match (use list_products with client-side filters) ❌ Product reviews (use list_standard_product_reviews or list_market_intelligence_product_reviews instead) ❌ Bulk product retrieval (use list_products with pagination) AVAILABLE FIELDS: name, domain, slug, image_url, detail_description, g2_url, public_detail_url, review_count, star_rating, type, write_review_url AVAILABLE RELATIONSHIPS: categories - Product category memberships discussions - Product discussion threads vendor - Product vendor information WHEN TO USE: - You have a product UUID or Slug from a previous API call - You need full product details for a specific product - You want to explore relationships (vendor, categories) - Building a product detail page or profile COMMON QUERY PATTERNS: 1. Full product details: fields="name,detail_description,star_rating,review_count,g2_url" 2. Product with vendor context: fields="name,domain,star_rating", include="vendor" 3. Product with category memberships: fields="name,slug", include="categories" 4. Complete product profile: (empty fields parameter), include="vendor,categories,discussions" PERFORMANCE CONSIDERATIONS: - Single product lookup (no pagination needed) - Empty fields parameter returns all available fields - Relationships add to response size (include only when needed) - Most efficient for UUID or Slug-based lookups
show_product_creation_form
ChatGPTDisplay the Prefab form for adding one product to G2.
show_product_csv_upload
ChatGPTDisplay the Prefab CSV picker and the product import template.
show_profile_section
ChatGPTBETA: profile editing is new and still in development. Show one section of a product's G2 profile: its current values and the list of fields the vendor may change (meta.fields). DESIGNED FOR: ✅ Reading a section before proposing a change to it ✅ Answering questions about the vendor's OWN product listing: its name, description, languages, hosted and on-premise options, logos and banner (use this, not show_product, for the vendor's own product) ✅ Checking which fields are editable and what values they accept NOT DESIGNED FOR: ❌ Changing anything (use propose_profile_change, then approve_profile_change) ❌ Reading other companies' products SECTIONS: information Product name, description, languages, hosted and on-premise deployment, product logo and Grid logo. The short description is read only (G2 Support changes it); meta.fields names where to send the vendor. The logo and the Grid logo are separate images: if the vendor may want the same file for both, offer it, and never change both without asking. banner The banner at the top of the G2 profile. Only on plans that include a banner; otherwise the call is refused. AGENT GUIDANCE: Each field in meta.fields has a description in plain words, plus its type, rules, options and whether it is editable. Use the descriptions when you explain a field to the user or when a request could mean more than one field. Only fields with editable: true can be proposed. The seller (company) name is not part of this section and cannot be changed with these tools; do not guess where it is changed.
show_research_board
ChatGPTDisplay an interactive research board with product comparisons.
show_review_campaign
ChatGPTRead-only lookup of an existing review campaign by id, for a product. Returns the same attribute shape as create_review_campaign, so a caller can re-check a campaign's state after creation instead of relying only on the one-time create response. DESIGNED FOR: ✅ Confirming a campaign was created with the expected attributes ✅ Checking whether a campaign is still actively collecting reviews (running) NOT DESIGNED FOR: ❌ Creating a campaign (use create_review_campaign) ❌ Listing all campaigns for a product — no tool exists for that yet ❌ Editing or pausing an existing campaign — no tool exists for that yet
show_review_campaign_budget
ChatGPTRead-only pre-flight check: call this BEFORE create_review_campaign whenever the vendor wants an incentivized campaign. Returns available G2 credit, whether an eligible default payment method is on file, whether External Invoice fallback is available, and how many reviews the available credit alone would cover at a given reward_amount. Funding is never an explicit choice between credit and a payment method: G2 credit is applied automatically to per-review charges, and the eligible payment method covers any remainder. The API uses the product's default payment method and can fall back to External Invoice when can_use_external_invoice=true. A product with neither is rejected even if its credit would cover the goal. Use this tool's response to tell the vendor what's available and, if both options are absent, that they need to add a payment method (a real card can only be added by the vendor themselves in a browser — no tool can do this on their behalf). DESIGNED FOR: ✅ Checking funding status before proposing an incentivized campaign ✅ Estimating how many reviews a given reward_amount would cover from credit alone (informational only — does not gate creation) NOT DESIGNED FOR: ❌ Adding or changing a payment method (there is no tool for this — direct the vendor to add_payment_method_url in a browser) ❌ Non-incentivized campaigns — skip this and go straight to create_review_campaign when incentivized=false
show_upgrade_widget
ChatGPTDisplay the G2 plan upgrade widget when a tool returns an access restriction error.
show_usability_report
ChatGPTDeprecated: use show_index_report with report_type="usability" instead. Render a licensed Usability Index report by report UUID. This compatibility tool remains available during the ChatGPT connector cutover; it requires reports.read and returns the same G2 Index widget and report snapshot.
show_vendor
ChatGPTRetrieve detailed information for a specific vendor by UUID. DESIGNED FOR: ✅ Single vendor detail retrieval with known vendor UUID ✅ Vendor profile pages and detail views ✅ Vendor metadata lookup (descriptions, websites, product counts) ✅ Product portfolio exploration for specific vendor NOT DESIGNED FOR: ❌ Vendor discovery (use list_vendors instead) ❌ Searching by name or partial match (use list_vendors with client-side filters) ❌ Product-level details (use list_products or show_product instead) ❌ Bulk vendor retrieval (use list_vendors with pagination) ACCESS SCOPE: Returns all approved vendors on G2. AVAILABLE FIELDS: name, description, company_website, slug, public_products_count, updated_at AVAILABLE RELATIONSHIPS: products - Vendor's product portfolio WHEN TO USE: - You have a vendor UUID from a previous API call - You need full vendor details for a specific vendor - You want to explore vendor's product portfolio - Building a vendor profile or detail page COMMON QUERY PATTERNS: 1. Full vendor details: fields="name,description,company_website,public_products_count" 2. Vendor with product portfolio: fields="name,slug", include="products" 3. Complete vendor profile: (empty fields parameter), include="products" 4. Basic vendor info: fields="name,company_website,slug" PERFORMANCE CONSIDERATIONS: - Single vendor lookup (no pagination needed) - Empty fields parameter returns all available fields - Products relationship can return large result set - Most efficient for UUID-based lookups
store_profile_upload
ChatGPTStore the image a vendor picked in the upload card. Only the card calls this tool; it is not meant for the model. It checks the file against the field's rules and returns the address of the stored file.
update_research_board
ChatGPTUpdate a research board's title. DESIGNED FOR: ✅ Renaming an existing research board NOT DESIGNED FOR: ❌ Creating new boards (use create_research_board) ❌ Adding or removing products (use add_products_to_research_board / remove_products_from_research_board)
validate_product_creation_step
ChatGPTInternal Prefab form callback. Validate one step of the guided product creation flow and return UI state only. Never creates a product or calls G2.