get_conversation_analysis
ChatGPTTool Description Batch AI analysis endpoint for Zoom Revenue Accelerator conversations. Accepts up to 5 conversation IDs per request and returns structured AI-generated analysis such as summary, sentiment, next steps, good questions, topics, indicators, deal memo, and speaker metrics. When to Use Use this tool when you need to retrieve AI-generated insights for one or more sales conversations, such as generating meeting summaries, evaluating sentiment, extracting action items, reviewing deal health signals, or analyzing speaker metrics. For raw spoken content, use get_conversation_transcript instead. Must Do - Always provide at least one conversation ID in conversation_ids - Explicitly provide analysis_types when you only need specific analysis sections, to reduce response payload and improve efficiency - If analysis_types is omitted or empty, all available analysis types are returned - Limit requests to a maximum of 5 conversation IDs; IDs beyond the 5th are silently ignored Error Handling - If a conversation is inaccessible or has no analysis data, its entry is still returned but with an empty analysis object — this is not an error - Absent fields in the analysis object indicate either the type was not requested or no data exists for that section - summary may be truncated to 5 KB (UTF-8); no trailing marker is added - indicators list is deduplicated by indicator ID Call Chain - UPSTREAM: Call search_conversations first to obtain conversation_ids - DOWNSTREAM: No further tools needed
get_conversation_analysis
ChatGPTTool Description Batch AI analysis endpoint for Zoom Revenue Accelerator conversations. Accepts up to 5 conversation IDs per request and returns structured AI-generated analysis such as summary, sentiment, next steps, good questions, topics, indicators, deal memo, and speaker metrics. When to Use Use this tool when you need to retrieve AI-generated insights for one or more sales conversations, such as generating meeting summaries, evaluating sentiment, extracting action items, reviewing deal health signals, or analyzing speaker metrics. For raw spoken content, use get_conversation_transcript instead. Must Do - Always provide at least one conversation ID in conversation_ids - Explicitly provide analysis_types when you only need specific analysis sections, to reduce response payload and improve efficiency - If analysis_types is omitted or empty, all available analysis types are returned - Limit requests to a maximum of 5 conversation IDs; IDs beyond the 5th are silently ignored Error Handling - If a conversation is inaccessible or has no analysis data, its entry is still returned but with an empty analysis object — this is not an error - Absent fields in the analysis object indicate either the type was not requested or no data exists for that section - summary may be truncated to 5 KB (UTF-8); no trailing marker is added - indicators list is deduplicated by indicator ID Call Chain - UPSTREAM: Call search_conversations first to obtain conversation_ids - DOWNSTREAM: No further tools needed
get_conversation_transcript
ChatGPTTool Description Retrieves the full transcript of a specific Zoom Revenue Accelerator conversation. Returns a compact token-optimized transcript string where each line is prefixed with a numeric speaker ID, plus a speaker_defs list that maps each ID to the speaker's display name and role. Optionally filters to only segments containing a keyword. When to Use Use this tool when you need access to the raw spoken content of a conversation — for example, to find specific quotes, verify what was said, or analyze dialogue in detail. For AI-extracted insights such as summary, action items, and sentiment, use get_conversation_analysis instead. Must Do - Always provide conversation_id (obtained from search_conversations) - Use keyword to narrow transcript output when looking for specific topics — this reduces token consumption significantly Error Handling - If the conversation has no transcript or no segments match the keyword, transcript returns an empty string "" - Speaker IDs in the transcript string are numeric (e.g., "1:", "2:") — use speaker_defs to resolve them to display names Call Chain - UPSTREAM: Call search_conversations first to obtain the conversation_id - DOWNSTREAM: No further tools needed
get_conversation_transcript
ChatGPTTool Description Retrieves the full transcript of a specific Zoom Revenue Accelerator conversation. Returns a compact token-optimized transcript string where each line is prefixed with a numeric speaker ID, plus a speaker_defs list that maps each ID to the speaker's display name and role. Optionally filters to only segments containing a keyword. When to Use Use this tool when you need access to the raw spoken content of a conversation — for example, to find specific quotes, verify what was said, or analyze dialogue in detail. For AI-extracted insights such as summary, action items, and sentiment, use get_conversation_analysis instead. Must Do - Always provide conversation_id (obtained from search_conversations) - Use keyword to narrow transcript output when looking for specific topics — this reduces token consumption significantly Error Handling - If the conversation has no transcript or no segments match the keyword, transcript returns an empty string "" - Speaker IDs in the transcript string are numeric (e.g., "1:", "2:") — use speaker_defs to resolve them to display names Call Chain - UPSTREAM: Call search_conversations first to obtain the conversation_id - DOWNSTREAM: No further tools needed
get_customer_accounts
ChatGPTTool Description Retrieves customer account records from Zoom Revenue Accelerator. Returns company profile details (industry, domain, website, employee count, annual revenue), CRM integration identifiers, ownership, and deal/contact counts. Supports lookup by account IDs or keyword search. When to Use Use this tool when the user refers to a company by name and you need the internal account_id — for example, before filtering conversations, deals, or contacts for a specific company. Also use it to retrieve general account profile information. Must Do - Use keyword to search by company name when you don't already have the account_id - Confirm the correct account with the user if the keyword returns multiple results with similar names Error Handling - Returns an empty accounts array when no accounts match the query - If no exact company match is found, try a shorter or less specific keyword Call Chain - UPSTREAM: No dependencies - DOWNSTREAM: account_id → get_customer_contacts (customer_account_id), search_deals (customer_account_id), search_conversations (customer_account_id)
get_customer_accounts
ChatGPTTool Description Retrieves customer account records from Zoom Revenue Accelerator. Returns company profile details (industry, domain, website, employee count, annual revenue), CRM integration identifiers, ownership, and deal/contact counts. Supports lookup by account IDs or keyword search. When to Use Use this tool when the user refers to a company by name and you need the internal account_id — for example, before filtering conversations, deals, or contacts for a specific company. Also use it to retrieve general account profile information. Must Do - Use keyword to search by company name when you don't already have the account_id - Confirm the correct account with the user if the keyword returns multiple results with similar names Error Handling - Returns an empty accounts array when no accounts match the query - If no exact company match is found, try a shorter or less specific keyword Call Chain - UPSTREAM: No dependencies - DOWNSTREAM: account_id → get_customer_contacts (customer_account_id), search_deals (customer_account_id), search_conversations (customer_account_id)
get_customer_contacts
ChatGPTTool Description Retrieves customer contact records from Zoom Revenue Accelerator. Returns contact profile (name, email, phone, job title), company affiliation, last engagement date, deal count, and optionally recent deals, recent meetings, and custom CRM fields. When to Use Use this tool when you need to look up external customer contacts — for example, to find a contact's ID before filtering conversations by that contact, or to review their recent deal and meeting history. Must Do - Use customer_account_id to scope the search when looking for contacts at a specific company (obtain it from get_customer_accounts first) - Set include_recent_deals=true only when deal history is specifically needed — it increases response payload - Set include_recent_meetings=true only when meeting history is specifically needed Error Handling - Returns an empty contacts array when no contacts match the filter - recent_deal is absent when the contact has no associated deals, or when include_recent_deals is false - recent_meetings is absent or empty when the contact has no associated meetings, or when include_recent_meetings is false - custom_data is absent or empty when no custom CRM data exists, or when include_custom_data is false Call Chain - UPSTREAM: get_customer_accounts → customer_account_id (scope to a company) - DOWNSTREAM: contact_id → search_conversations (customer_contact_ids)
get_customer_contacts
ChatGPTTool Description Retrieves customer contact records from Zoom Revenue Accelerator. Returns contact profile (name, email, phone, job title), company affiliation, last engagement date, deal count, and optionally recent deals, recent meetings, and custom CRM fields. When to Use Use this tool when you need to look up external customer contacts — for example, to find a contact's ID before filtering conversations by that contact, or to review their recent deal and meeting history. Must Do - Use customer_account_id to scope the search when looking for contacts at a specific company (obtain it from get_customer_accounts first) - Set include_recent_deals=true only when deal history is specifically needed — it increases response payload - Set include_recent_meetings=true only when meeting history is specifically needed Error Handling - Returns an empty contacts array when no contacts match the filter - recent_deal is absent when the contact has no associated deals, or when include_recent_deals is false - recent_meetings is absent or empty when the contact has no associated meetings, or when include_recent_meetings is false - custom_data is absent or empty when no custom CRM data exists, or when include_custom_data is false Call Chain - UPSTREAM: get_customer_accounts → customer_account_id (scope to a company) - DOWNSTREAM: contact_id → search_conversations (customer_contact_ids)
get_deal_activities_v2
ChatGPTTool Description Returns a paginated list of activities (meetings, calls, emails) logged against a specific CRM deal. Supports filtering by activity type, date range, and optional content inclusion. Also returns participants and the deal stage at the time of each activity. When to Use Use this tool when the user wants to review the interaction history for a specific deal — for example, how many calls have been made, what was discussed in recent meetings, or who participated in deal activities. Must Do - Always provide deal_id (obtained from search_deals) - Set include_content=true only when activity content (call summary / meeting description) is specifically needed — it increases response payload - For best performance, keep date range within one month Error Handling - Returns an empty data_items array when no activities match the filters - next_page_token is empty when there are no more pages Call Chain - UPSTREAM: Call search_deals first to obtain the deal_id - DOWNSTREAM: No further tools needed
get_deal_activities_v2
ChatGPTTool Description Returns a paginated list of activities (meetings, calls, emails) logged against a specific CRM deal. Supports filtering by activity type, date range, and optional content inclusion. Also returns participants and the deal stage at the time of each activity. When to Use Use this tool when the user wants to review the interaction history for a specific deal — for example, how many calls have been made, what was discussed in recent meetings, or who participated in deal activities. Must Do - Always provide deal_id (obtained from search_deals) - Set include_content=true only when activity content (call summary / meeting description) is specifically needed — it increases response payload - For best performance, keep date range within one month Error Handling - Returns an empty data_items array when no activities match the filters - next_page_token is empty when there are no more pages Call Chain - UPSTREAM: Call search_deals first to obtain the deal_id - DOWNSTREAM: No further tools needed
get_deal_analysis
ChatGPTTool Description Performs AI-driven analysis on a specific CRM deal. Returns a STORY (narrative summary of the deal's progression and current status) and/or a PLAYBOOK evaluation (per-element compliance notes against configured sales methodologies such as MEDDIC). Both types can be requested in a single call. When to Use Use this tool when you need to understand the overall health, momentum, and risks of a specific deal, or to evaluate the deal against a sales playbook methodology. Call get_deal_detail_v2 first if you also need structured deal data such as stage, amount, or contacts. Must Do - Always provide deal_id (obtained from search_deals) - Set analysis_type to STORY, PLAYBOOK, or ALL (default when omitted) Error Handling - If the deal has no analysis data for the requested type, the corresponding field (deal_story or playbook_items) is absent from the response — this is expected for newly created deals or deals without sufficient activity - playbook_items is absent when no playbook is configured for the deal Call Chain - UPSTREAM: Call search_deals first to obtain the deal_id - DOWNSTREAM: No further tools needed
get_deal_analysis
ChatGPTTool Description Performs AI-driven analysis on a specific CRM deal. Returns a STORY (narrative summary of the deal's progression and current status) and/or a PLAYBOOK evaluation (per-element compliance notes against configured sales methodologies such as MEDDIC). Both types can be requested in a single call. When to Use Use this tool when you need to understand the overall health, momentum, and risks of a specific deal, or to evaluate the deal against a sales playbook methodology. Call get_deal_detail_v2 first if you also need structured deal data such as stage, amount, or contacts. Must Do - Always provide deal_id (obtained from search_deals) - Set analysis_type to STORY, PLAYBOOK, or ALL (default when omitted) Error Handling - If the deal has no analysis data for the requested type, the corresponding field (deal_story or playbook_items) is absent from the response — this is expected for newly created deals or deals without sufficient activity - playbook_items is absent when no playbook is configured for the deal Call Chain - UPSTREAM: Call search_deals first to obtain the deal_id - DOWNSTREAM: No further tools needed
get_deal_detail_v2
ChatGPTTool Description Batch-fetches full deal details for up to 20 deal IDs. Returns comprehensive deal data including pipeline stage, value, owner, customer account, activity count, contact count, trend indicators, deal team members, key contacts, and optionally custom CRM fields. When to Use Use this tool after search_deals when you need complete deal information beyond the summary fields — for example, to review team composition, key contacts, or custom CRM fields for deal preparation or board review. Must Do - Always provide deal_ids (obtained from search_deals); maximum 20 per request - Set include_custom_fields=true only when CRM custom fields are specifically needed — it increases response payload Error Handling - Returns only the deals the caller has access to; inaccessible deal IDs are silently excluded - total in the response reflects the number of deals actually returned, not the number of IDs requested Call Chain - UPSTREAM: Call search_deals first to obtain the deal_ids - DOWNSTREAM: customer_account_id → get_customer_accounts; deal_id → get_deal_analysis, get_deal_activities_v2
get_deal_detail_v2
ChatGPTTool Description Batch-fetches full deal details for up to 20 deal IDs. Returns comprehensive deal data including pipeline stage, value, owner, customer account, activity count, contact count, trend indicators, deal team members, key contacts, and optionally custom CRM fields. When to Use Use this tool after search_deals when you need complete deal information beyond the summary fields — for example, to review team composition, key contacts, or custom CRM fields for deal preparation or board review. Must Do - Always provide deal_ids (obtained from search_deals); maximum 20 per request - Set include_custom_fields=true only when CRM custom fields are specifically needed — it increases response payload Error Handling - Returns only the deals the caller has access to; inaccessible deal IDs are silently excluded - total in the response reflects the number of deals actually returned, not the number of IDs requested Call Chain - UPSTREAM: Call search_deals first to obtain the deal_ids - DOWNSTREAM: customer_account_id → get_customer_accounts; deal_id → get_deal_analysis, get_deal_activities_v2
get_deal_stages
ChatGPTTool Description Retrieves the list of CRM deal pipeline stages configured for this account. Returns each stage's display name, sort order, and flags indicating whether it is a closed stage and whether it represents a won deal. When to Use Use this tool before filtering deals by stage (search_deals → deal_stages), or when the user asks about available pipeline stages. Call it once at the start of a deal workflow to obtain valid stage names. Must Do - Use the exact stage names from the response when passing to search_deals Error Handling - Returns an empty stages list when no stages are configured for the account Call Chain - UPSTREAM: No dependencies - DOWNSTREAM: stage names → search_deals (deal_stages)
get_deal_stages
ChatGPTTool Description Retrieves the list of CRM deal pipeline stages configured for this account. Returns each stage's display name, sort order, and flags indicating whether it is a closed stage and whether it represents a won deal. When to Use Use this tool before filtering deals by stage (search_deals → deal_stages), or when the user asks about available pipeline stages. Call it once at the start of a deal workflow to obtain valid stage names. Must Do - Use the exact stage names from the response when passing to search_deals Error Handling - Returns an empty stages list when no stages are configured for the account Call Chain - UPSTREAM: No dependencies - DOWNSTREAM: stage names → search_deals (deal_stages)
get_manager_team_and_member
ChatGPTTool Description Retrieves team information in Zoom Revenue Accelerator. When called without team_ids, it returns teams managed by the current user together with their managers. When called with team_ids, it returns detailed recursive member lists for the specified teams. When to Use Use this tool ONLY when: - The user asks about team structure, team members, or team managers - search_conversations or search_deals needs team_ids for scope=SPECIFIC — i.e., the user names a specific team Do NOT use this tool when scope=MY_MANAGED — that scope does not require team_ids Must Do - Start with overview mode (omit team_ids) to discover teams managed by the current user - Use detail mode (provide team_ids) only when you need the full member list for one or more specific teams - members is capped at 200 entries per team in detail mode Error Handling - Returns an empty items array when the current user manages no teams - members is only present in detail mode Call Chain - UPSTREAM: No dependencies - DOWNSTREAM (only when scope=SPECIFIC): team_id -> search_conversations (team_ids), search_deals (team_ids)
get_manager_team_and_member
ChatGPTTool Description Retrieves team information in Zoom Revenue Accelerator. When called without team_ids, it returns teams managed by the current user together with their managers. When called with team_ids, it returns detailed recursive member lists for the specified teams. When to Use Use this tool ONLY when: - The user asks about team structure, team members, or team managers - search_conversations or search_deals needs team_ids for scope=SPECIFIC — i.e., the user names a specific team Do NOT use this tool when scope=MY_MANAGED — that scope does not require team_ids Must Do - Start with overview mode (omit team_ids) to discover teams managed by the current user - Use detail mode (provide team_ids) only when you need the full member list for one or more specific teams - members is capped at 200 entries per team in detail mode Error Handling - Returns an empty items array when the current user manages no teams - members is only present in detail mode Call Chain - UPSTREAM: No dependencies - DOWNSTREAM (only when scope=SPECIFIC): team_id -> search_conversations (team_ids), search_deals (team_ids)
get_scorecard_sessions
ChatGPTTool Description Retrieves scorecard evaluation sessions for a list of Zoom Revenue Accelerator conversations. Returns per-session score, scorecard template name, ratee (evaluated agent) info, and flattened score items with individual scores, selected answers, and AI justifications. When to Use Use this tool to check whether specific conversations have been evaluated by a QM scorecard and to retrieve the results. Typical use: after fetching conversations via search_conversations, pass their IDs here to retrieve coaching or quality evaluation sessions. Must Do - Always provide conversation_ids — omitting it returns an empty result in the current version - Limit to a maximum of 20 conversation IDs per request Error Handling - Conversations that have no scorecard session are simply absent from the returned sessions list — this is expected and not an error - score_percent is calculated as (score / total_score) * 100 Call Chain - UPSTREAM: Call search_conversations first to obtain the conversation_ids (up to 20) - DOWNSTREAM: No further tools needed
get_scorecard_sessions
ChatGPTTool Description Retrieves scorecard evaluation sessions for a list of Zoom Revenue Accelerator conversations. Returns per-session score, scorecard template name, ratee (evaluated agent) info, and flattened score items with individual scores, selected answers, and AI justifications. When to Use Use this tool to check whether specific conversations have been evaluated by a QM scorecard and to retrieve the results. Typical use: after fetching conversations via search_conversations, pass their IDs here to retrieve coaching or quality evaluation sessions. Must Do - Always provide conversation_ids — omitting it returns an empty result in the current version - Limit to a maximum of 20 conversation IDs per request Error Handling - Conversations that have no scorecard session are simply absent from the returned sessions list — this is expected and not an error - score_percent is calculated as (score / total_score) * 100 Call Chain - UPSTREAM: Call search_conversations first to obtain the conversation_ids (up to 20) - DOWNSTREAM: No further tools needed
search_conversations
ChatGPTTool Description Retrieves a paginated list of Zoom Revenue Accelerator conversations (meetings and phone calls) with flexible filtering. Returns conversation IDs, host info, participants, duration, and linked deal and team metadata. When to Use Use this tool to discover conversations before drilling into their content. It is the primary entry point for conversation workflows. Typical user intents: - "Show me my recent meetings" → scope=MINE - "Find calls where competitor X was mentioned" → resolve indicator first via search_indicators, then filter by indicator_id - "What conversations did Alice have last week?" → resolve person first, then filter by internal_user_ids or customer_contact_ids - "List conversations linked to the Acme deal" → resolve deal first via search_deals, then filter by deal_ids Key Parameters | Parameter | Type | Notes | |---|---|---| | scope | string | MINE, MY_MANAGED, SPECIFIC, ALL | | conversation_type | string | meeting, phone, call_queue, all. Only pass when the user explicitly specifies a type; otherwise omit entirely | | indicator_id | string | Single ID from search_indicators (NOT an array) | | keyword | string | Searches conversation title only, NOT transcript content | | internal_user_ids | string[] | From search_internal_users | | team_ids | string[] | Only when scope=SPECIFIC | | deal_ids | string[] | From search_deals | | customer_contact_ids | string[] | From get_customer_contacts | | customer_account_id | string | From get_customer_accounts | | return_conversation_id_only | boolean | Optimize for ID-only downstream workflows | Must Do Scope - If scope is ambiguous, always ask the user to choose — do NOT guess - scope=MY_MANAGED does NOT require get_manager_team_and_member - Use team_ids only with scope=SPECIFIC Conversation Type - Valid values: meeting, phone, call_queue, all (lowercase) - Only pass `conversation_type` when the user explicitly requests a specific type (e.g., "show me only meetings" → meeting, "show me phone calls" → phone) - If the user does not specify a type, do NOT pass this parameter — let the server return all types by default - Do NOT infer conversation_type from casual wording like "Find calls where…" — treat "calls" as a generic term for conversations unless the user clearly means phone calls only Person Resolution - If unclear whether internal or external, call search_internal_users and get_customer_contacts in parallel - If both return matches, ask user to confirm Keyword vs Indicator - keyword = conversation title search; indicator_id = content detection - If ambiguous, ask user to clarify - Do NOT copy indicator name into keyword after resolving indicator_id - If search_indicators returns no match, inform user and ask for next step Date Range - Default window: last 30 days - Max span: 3 years per query - `date_range` MUST be a JSON object, NOT a comma-separated string - Correct: {"from": "2026-06-18", "to": "2026-06-25"} - Wrong: "2026-06-18,2026-06-25" - If providing date_range, both from and to are required and must be in yyyy-MM-dd format Must NOT Do - Do NOT pass indicator_id as an array - Do NOT auto-populate keyword from indicator name - Do NOT broaden scope automatically when no results found - Do NOT guess scope when unclear - Do NOT pass conversation_type unless the user explicitly specifies a conversation type Error Handling - Empty conversations array when no matches - participants capped at first 20 (internal users prioritized) - next_page_token is empty when no more pages Call Chain - UPSTREAM: search_indicators → indicator_id; search_internal_users → internal_user_ids; search_deals → deal_ids; get_customer_accounts → customer_account_id; get_customer_contacts → customer_contact_ids; get_manager_team_and_member → team_ids (only for scope=SPECIFIC) - DOWNSTREAM: get_conversation_analysis, get_conversation_transcript, `ge…
search_conversations
ChatGPTTool Description Retrieves a paginated list of Zoom Revenue Accelerator conversations (meetings and phone calls) with flexible filtering. Returns conversation IDs, host info, participants, duration, and linked deal and team metadata. When to Use Use this tool to discover conversations before drilling into their content. It is the primary entry point for conversation workflows. Typical user intents: - "Show me my recent meetings" → scope=MINE - "Find calls where competitor X was mentioned" → resolve indicator first via search_indicators, then filter by indicator_id - "What conversations did Alice have last week?" → resolve person first, then filter by internal_user_ids or customer_contact_ids - "List conversations linked to the Acme deal" → resolve deal first via search_deals, then filter by deal_ids Key Parameters | Parameter | Type | Notes | |---|---|---| | scope | string | MINE, MY_MANAGED, SPECIFIC, ALL | | conversation_type | string | meeting, phone, call_queue, all. Only pass when the user explicitly specifies a type; otherwise omit entirely | | indicator_id | string | Single ID from search_indicators (NOT an array) | | keyword | string | Searches conversation title only, NOT transcript content | | internal_user_ids | string[] | From search_internal_users | | team_ids | string[] | Only when scope=SPECIFIC | | deal_ids | string[] | From search_deals | | customer_contact_ids | string[] | From get_customer_contacts | | customer_account_id | string | From get_customer_accounts | | return_conversation_id_only | boolean | Optimize for ID-only downstream workflows | Must Do Scope - If scope is ambiguous, always ask the user to choose — do NOT guess - scope=MY_MANAGED does NOT require get_manager_team_and_member - Use team_ids only with scope=SPECIFIC Conversation Type - Valid values: meeting, phone, call_queue, all (lowercase) - Only pass `conversation_type` when the user explicitly requests a specific type (e.g., "show me only meetings" → meeting, "show me phone calls" → phone) - If the user does not specify a type, do NOT pass this parameter — let the server return all types by default - Do NOT infer conversation_type from casual wording like "Find calls where…" — treat "calls" as a generic term for conversations unless the user clearly means phone calls only Person Resolution - If unclear whether internal or external, call search_internal_users and get_customer_contacts in parallel - If both return matches, ask user to confirm Keyword vs Indicator - keyword = conversation title search; indicator_id = content detection - If ambiguous, ask user to clarify - Do NOT copy indicator name into keyword after resolving indicator_id - If search_indicators returns no match, inform user and ask for next step Date Range - Default window: last 30 days - Max span: 3 years per query - `date_range` MUST be a JSON object, NOT a comma-separated string - Correct: {"from": "2026-06-18", "to": "2026-06-25"} - Wrong: "2026-06-18,2026-06-25" - If providing date_range, both from and to are required and must be in yyyy-MM-dd format Must NOT Do - Do NOT pass indicator_id as an array - Do NOT auto-populate keyword from indicator name - Do NOT broaden scope automatically when no results found - Do NOT guess scope when unclear - Do NOT pass conversation_type unless the user explicitly specifies a conversation type Error Handling - Empty conversations array when no matches - participants capped at first 20 (internal users prioritized) - next_page_token is empty when no more pages Call Chain - UPSTREAM: search_indicators → indicator_id; search_internal_users → internal_user_ids; search_deals → deal_ids; get_customer_accounts → customer_account_id; get_customer_contacts → customer_contact_ids; get_manager_team_and_member → team_ids (only for scope=SPECIFIC) - DOWNSTREAM: get_conversation_analysis, get_conversation_transcript, `ge…
search_deals
ChatGPTTool Description Searches CRM deals in Zoom Revenue Accelerator with flexible filtering by scope, keyword, pipeline stage, owner, amount range, close date range, last activity date range, customer account, and team membership. Returns paginated deal summaries including stage, amount, trend indicators, owner info, customer account, and next step. When to Use Use this tool when the user wants to find, filter, or list deals, for example "show me open deals closing this quarter" or "find deals owned by Alice". It is the primary entry point for deal workflows: first retrieve matching deals here, then pass their IDs to get_deal_detail_v2, get_deal_analysis, get_deal_activities_v2, or use them in search_conversations. Must Do - If the user's requested ownership scope is ambiguous, ask them to choose among MINE, MY_MANAGED, SPECIFIC, or ALL - When the user wants deals from their managed teams, set scope=MY_MANAGED directly — do NOT call get_manager_team_and_member first - Use team_ids only when scope is SPECIFIC - Only call get_manager_team_and_member when scope is SPECIFIC and the user names a specific team that needs to be resolved to team_ids - Resolve stage names with get_deal_stages before passing values to deal_stages - Resolve person names with search_internal_users before passing values to deal_owner_ids or deal_team_member_ids - Resolve company names with get_customer_accounts before passing values to customer_account_id - When no date filter is provided, the server defaults to the last 90 days of activity; mention the effective time window in your response if it matters to the answer - Always check and mention the effective has_activities behavior: - If has_activities=true, explain that the results include only deals with activity records - If has_activities=false, explain that the results include deals regardless of activity records Error Handling - Returns an empty data_items array when no deals match the filters - Do not broaden scope automatically when no results are found; confirm with the user first - The effective date window applied is reflected in expected_close_date_range and last_activity_date_range - next_page_token is empty when there are no more pages Call Chain - UPSTREAM (only when scope=SPECIFIC): get_manager_team_and_member -> team_ids; get_deal_stages -> deal_stages; search_internal_users -> deal_owner_ids / deal_team_member_ids; get_customer_accounts -> customer_account_id - DOWNSTREAM: get_deal_detail_v2, get_deal_analysis, get_deal_activities_v2, search_conversations
search_deals
ChatGPTTool Description Searches CRM deals in Zoom Revenue Accelerator with flexible filtering by scope, keyword, pipeline stage, owner, amount range, close date range, last activity date range, customer account, and team membership. Returns paginated deal summaries including stage, amount, trend indicators, owner info, customer account, and next step. When to Use Use this tool when the user wants to find, filter, or list deals, for example "show me open deals closing this quarter" or "find deals owned by Alice". It is the primary entry point for deal workflows: first retrieve matching deals here, then pass their IDs to get_deal_detail_v2, get_deal_analysis, get_deal_activities_v2, or use them in search_conversations. Must Do - If the user's requested ownership scope is ambiguous, ask them to choose among MINE, MY_MANAGED, SPECIFIC, or ALL - When the user wants deals from their managed teams, set scope=MY_MANAGED directly — do NOT call get_manager_team_and_member first - Use team_ids only when scope is SPECIFIC - Only call get_manager_team_and_member when scope is SPECIFIC and the user names a specific team that needs to be resolved to team_ids - Resolve stage names with get_deal_stages before passing values to deal_stages - Resolve person names with search_internal_users before passing values to deal_owner_ids or deal_team_member_ids - Resolve company names with get_customer_accounts before passing values to customer_account_id - When no date filter is provided, the server defaults to the last 90 days of activity; mention the effective time window in your response if it matters to the answer - Always check and mention the effective has_activities behavior: - If has_activities=true, explain that the results include only deals with activity records - If has_activities=false, explain that the results include deals regardless of activity records Error Handling - Returns an empty data_items array when no deals match the filters - Do not broaden scope automatically when no results are found; confirm with the user first - The effective date window applied is reflected in expected_close_date_range and last_activity_date_range - next_page_token is empty when there are no more pages Call Chain - UPSTREAM (only when scope=SPECIFIC): get_manager_team_and_member -> team_ids; get_deal_stages -> deal_stages; search_internal_users -> deal_owner_ids / deal_team_member_ids; get_customer_accounts -> customer_account_id - DOWNSTREAM: get_deal_detail_v2, get_deal_analysis, get_deal_activities_v2, search_conversations
search_indicators
ChatGPTTool Description Returns the list of keyword/topic indicator trackers configured by the account admin. Each indicator monitors whether specific words, phrases, or topics were mentioned in a conversation. Detection types: - keyword_basic — any keyword in trigger_keywords appears in the transcript - keyword_advanced — boolean expression (AND / OR / NOT) over terms - guiding_sentences — AI semantic similarity match; no keywords exposed - keyword_theme — multiple named keyword sub-groups; fires when any group matches When to Use - User mentions an indicator by name → resolve to indicator_id, then pass to search_conversations - User asks what indicators are configured → call with optional keyword Must Do - Always use keyword to narrow results — never retrieve all indicators and silently pick one - keyword performs a contains-match on indicator name (case-insensitive) - If multiple matches are returned, present the full list and ask user to confirm - Default max_results is 20; do not increase unless user explicitly requests more Error Handling - If keyword returns no results, suggest widening the search term or ask user to verify the indicator name Call Chain - UPSTREAM: No prerequisites - DOWNSTREAM: indicator_id → search_conversations (indicator_id filter)
search_indicators
ChatGPTTool Description Returns the list of keyword/topic indicator trackers configured by the account admin. Each indicator monitors whether specific words, phrases, or topics were mentioned in a conversation. Detection types: - keyword_basic — any keyword in trigger_keywords appears in the transcript - keyword_advanced — boolean expression (AND / OR / NOT) over terms - guiding_sentences — AI semantic similarity match; no keywords exposed - keyword_theme — multiple named keyword sub-groups; fires when any group matches When to Use - User mentions an indicator by name → resolve to indicator_id, then pass to search_conversations - User asks what indicators are configured → call with optional keyword Must Do - Always use keyword to narrow results — never retrieve all indicators and silently pick one - keyword performs a contains-match on indicator name (case-insensitive) - If multiple matches are returned, present the full list and ask user to confirm - Default max_results is 20; do not increase unless user explicitly requests more Error Handling - If keyword returns no results, suggest widening the search term or ask user to verify the indicator name Call Chain - UPSTREAM: No prerequisites - DOWNSTREAM: indicator_id → search_conversations (indicator_id filter)
search_internal_users
ChatGPTTool Description Searches internal Zoom Revenue Accelerator users by display name or email fragment. Returns matching users with their user_id, display_name, and email. Use this tool to resolve a person's name into the internal user ID required by downstream filters. When to Use Use this tool whenever the user refers to a specific internal person by name, for example before filtering conversations by participant or filtering deals by owner or team member. Must Do - Use keyword to search by name or email fragment - Confirm with the user if multiple plausible results are returned and the intended person is ambiguous Error Handling - Returns an empty data_items array when no users match the keyword - Results may be limited; narrow the keyword if too many matches are returned Call Chain - UPSTREAM: No dependencies - DOWNSTREAM: user_id -> search_conversations (internal_user_ids), search_deals (deal_owner_ids, deal_team_member_ids)
search_internal_users
ChatGPTTool Description Searches internal Zoom Revenue Accelerator users by display name or email fragment. Returns matching users with their user_id, display_name, and email. Use this tool to resolve a person's name into the internal user ID required by downstream filters. When to Use Use this tool whenever the user refers to a specific internal person by name, for example before filtering conversations by participant or filtering deals by owner or team member. Must Do - Use keyword to search by name or email fragment - Confirm with the user if multiple plausible results are returned and the intended person is ambiguous Error Handling - Returns an empty data_items array when no users match the keyword - Results may be limited; narrow the keyword if too many matches are returned Call Chain - UPSTREAM: No dependencies - DOWNSTREAM: user_id -> search_conversations (internal_user_ids), search_deals (deal_owner_ids, deal_team_member_ids)
