aggregation_tool
ChatGPT<summary>Get aggregated data by natural language query for accounts, organizations, or enabled Custom Objects. When custom_object is set, queries the CO materialized table (e.g. invoices_materialized) with admin-curated schema. entity_type indicates which customer link column to use for scoping. Set aggregate_on='facts' with a fact_type to aggregate cross-entity AI fact analyses attached to the entity — counts, breakdowns, and filters over those analyses. Fact aggregation can filter and group by standard entity attributes AND the fact's fields together. If the given fact_type isn't available for the tenant, the tool returns the list of valid fact types.</summary> <returns> <description>Aggregated results (counts, sums, breakdowns by attribute or fact field, etc.)</description> </returns>
analyse_customer_interactions_by_topic
ChatGPTReturns topic LABELS with statistics: interaction counts, channel breakdown, and involved entities.
entity_ranker
ChatGPT<summary>Get top entities by metric, optionally filter by owner, entity IDs or custom query. Use metric=mrr, order=desc, limit=5 if these settings are not specified in the user query.</summary> <returns> <description>Explanation and list of entity IDs, names, and metric values</description> </returns>
fetch_entity_attribute_history
ChatGPTFetch historical values of an attribute or fact for the given entity. DISCOVERY PATTERN (ALWAYS FOLLOW THIS): When a user asks for history of any attribute: Step 1: Call fetch_entity_fact_sheet to see the current structure Step 2: Determine which section contains the attribute: - If in 'facts' section → use attribute_source='fact', attribute_name='<fact_type_key>' - If in 'custom' section → use attribute_source='custom', attribute_name='<custom_field_name>' - If in 'standard' section → standard history not supported Step 3: Call THIS tool with the correct parameters This pattern works for ANY fact_type (risk, health, engagement, etc.) - new fact types appear in the facts section automatically without code changes. PARAMETER DETAILS:
fetch_entity_fact_sheet
ChatGPT<summary>Fetch key facts about an account or organization by type and ID.</summary> <returns> <description>Key facts dictionary including headline report link with recent activities</description> </returns>
resolve_customer_team_member
ChatGPT<summary>Find user ID of customer team member (account owner, specialist, etc.) from name/partial name.</summary> <returns> <description>User ID and name, or multiple candidates if ambiguous</description> </returns>
resolve_entity
ChatGPT<summary>Find entity ID from name/partial name.</summary> <returns> <description>ID and name, or multiple matches if ambiguous</description> </returns>
search_customer_communications
ChatGPT<summary>Search and retrieve historical CSM-customer communications across multiple channels. Queries communication records (email, slack, notes) between Customer Success Managers and customers. Support tickets are excluded from results. At least one entity ID must be provided to constrain the search.</summary> <returns> <description>Grouped communication summaries per entity, newest first</description> </returns>
search_entity_assessments
ChatGPT<summary>Fetch recent health assessments (CSM Pulse, AI Pulse, notes) for given entities(accounts/organizations). Note: For ai_pulse history related fields, use this tool.</summary> <returns> <description>Grouped observation summaries per entity, newest first</description> </returns>
search_support_tickets
ChatGPT<summary>Search support ticket communications for the given entities.</summary> <returns> <description>Ticket summaries per entity, newest first</description> </returns>
