get_agent_decisions
ChatGPTRetrieve strategic recommendations for a single account, including an account summary, a recommended outreach strategy with value hypotheses, and a contact strategy with prioritized contacts to reach out to. Use this when you need actionable next steps for engaging an account.
get_agent_decisions_sections
ChatGPTRetrieve only specific sections of an account's strategic recommendations. Use this instead of get_agent_decisions when you only need one or two sections to keep the response smaller.
get_agent_memory
ChatGPTRetrieve the research profile for a single account. When summary_only=false (default), returns the full profile: { "report": str, "report_references": [ref, ...], "smart_fields": {label: {answer, reasoning, citations} | None}, "account_details": [{fact, citations}, ...], "pain_points": [{fact, citations}, ...], "products": [{fact, citations}, ...], "timing_signals": [{fact, citations}, ...], "contacts": [{fact, citations}, ...], "requirements": [{fact, citations}, ...], "initiatives": [{fact, citations}, ...], "other": [{fact, citations}, ...] } When summary_only=true, see the summary_only parameter description for the reduced shape.
get_agent_memory_sections
ChatGPTRetrieve specific sections of an account's research profile. Use this when you need to dive deeper into a specific section of the account's research profile. Returns a JSON object with only the requested sections.
get_bulk_agent_decisions
ChatGPTRetrieve strategic recommendations for multiple accounts in a single call. Maximum 10 accounts per request. Use this instead of calling get_agent_decisions repeatedly. Returns a JSON object with two fields: - "results": a map from account ID to that account's recommendations (only accounts with available decisioning are included) - "missing_account_ids": a list of requested IDs that had no decisioning (either the account does not exist in this tenant, or its decisioning has not yet been computed)
get_bulk_agent_memory
ChatGPTRetrieve research profiles for multiple accounts in a single call. Maximum 10 accounts per request. Use this instead of calling get_agent_memory repeatedly when you need data for more than one account. Returns a JSON object with two fields: - "results": a map from account ID to that account's research profile. Shape depends on summary_only (see below). Only accounts with available data are included. - "missing_account_ids": a list of requested IDs that had no research profile (either the account does not exist in this tenant, or its memory has not been computed yet) summary_only=true (the default) — each per-account value is: { "report": str, intelligence brief (truncated, may end with "…") "report_references": [ref], capped to the first few items "section_counts": { length of each per-section list "account_details": int, "initiatives": int, "pain_points": int, "products": int, "timing_signals": int, "contacts": int, "requirements": int, "other": int }, "smart_fields": {label: {"answer": str_or_int_or_dict} | None} in summary mode the reasoning + citations on smart fields are dropped } summary_only=false — each per-account value is the full profile: { "report": str, "report_references": [ref, ...], "smart_fields": {label: {"answer", "reasoning", "citations": [...]} | None}, "account_details": [{fact, citations}, ...], "pain_points": [{fact, citations}, ...], "products": [{fact, citations}, ...], "timing_signals": [{fact, citations}, ...], "contacts": [{fact, citations}, ...], "requirements": [{fact, citations}, ...], "initiatives": [{fact, citations}, ...], "other": [{fact, citations}, ...] } Full payloads can run into the hundreds of thousands of characters at the maximum account count — only pass summary_only=false when you need the per-section item bodies.
get_recent_account_activity
ChatGPTReturn account facts whose citations are dated within the last N days, sorted most-recent-first. Prefer this over get_agent_memory for deal-progression or 'what is happening right now' questions — it surfaces the freshest signals without requiring you to filter a large memory blob yourself. get_agent_decisions is a periodically-refreshed summary and may lag real activity by days; use this tool when recency matters.
get_workflow_instructions
ChatGPTGet the full instructions/template for a specific workflow by its label. Use list_workflows first to see available workflow labels. Returns the complete prompt template that guides how to execute the workflow.
list_workflows
ChatGPTList all available workflows/skills for you and your organization. Returns each workflow's label, display name, description of when to use it, and whether it's an org-level or personal workflow. Use this to discover what workflows are available before fetching specific instructions.
lookup_accounts
ChatGPTLook up accounts by name or Salesforce account ID. Supports fuzzy name matching. Use this to resolve an account name to its Salesforce ID, or look up the name for a known ID. Returns a list of matching accounts with their IDs and names.
generate_smart_fields
Claudeget_agent_decisions
Claudeget_agent_decisions_sections
Claudeget_agent_memory
Claudeget_agent_memory_sections
Claudeget_bulk_agent_decisions
Claudeget_bulk_agent_memory
Claudeget_recent_account_activity
Claudeget_smart_field_schemas
Claudeget_workflow_instructions
Claudelist_workflows
Claudelookup_accounts
Claudesearch_agent_memory
Claude
