Zentrik

Use this when the user asks to save one coherent customer-evidence source in Zentrik. evidenceType is the source format; context is the interaction or purpose. Account IDs must already exist. Returns a retry-safe confirmation with processing status.

create_evidence_record

Use this when ChatGPT needs the full Zentrik context for an item returned by search. Returns standard fetch content with evidence and relationship metadata.

fetch

Use this when the user asks whether a Zentrik evidence record created through ChatGPT has finished processing.

get_evidence_processing_status

Use this when the user asks why one already-identified Zentrik opportunity exists or what evidence supports that specific opportunity. Returns evidence-backed context, linked findings, and relationship counts. For rankings or comparisons among product options, use prepare_decision_review instead.

get_opportunity_evidence

Use this when ChatGPT needs to find a named Zentrik record or retrieve company-knowledge matches across customer evidence, findings, opportunities, or ideas. Returns standard search results that can be passed to fetch. For requests to decide what to build next, rank priorities, compare product bets, or open an interactive decision review, use prepare_decision_review instead.

search

Use this when ChatGPT needs evidence records or findings for one customer problem, product area, account, or source example. Returns compact records with evidence counts for ChatGPT summaries. For ranking or comparing what to build next, use prepare_decision_review instead.

search_evidence

Use this when the user asks what customer problems repeat across a topic in Zentrik. Searches evidence-backed findings so ChatGPT can summarize patterns and call out thin evidence.

summarize_repeated_problems