Docs

What the Growth Analyst Will and Won't Do

Evidence-based answers, explicit data gaps, and recommended next actions grounded in your records.

What Growth Analyst Is

Growth Analyst is an AI agent that answers revenue operations questions by querying your actual data — leads, conversations, pipeline, campaigns, revenue, integrations — and returning evidence-based answers with explicit citations.

**It is:** A reasoning layer over your operational data. **It is not:** A crystal ball, a data generator, or a replacement for human judgment.

Questions It Answers

| Category | Example Questions | |----------|-------------------| | Pipeline | "Which stages have the highest drop-off this quarter?" | | Leads | "What's our hot lead response time by rep?" | | Attribution | "Which campaigns produced the most qualified leads last month?" | | Conversations | "What are the top objection themes in closed-lost deals?" | | Revenue | "What's our CAC payback by channel?" | | Risks | "Which deals have been in proposal > 14 days?" | | Integrations | "Which Meta ad accounts haven't synced in 48 hrs?" | | Team | "Who has the most at-risk deals?" |

Evidence-Based Answers

Every answer includes: - **Citations** — Exact records queried (lead IDs, deal IDs, conversation IDs, campaign IDs) - **Data freshness** — When each source was last synced - **Confidence** — High / Medium / Low based on data completeness - **Gaps** — Explicit statement of what data is missing

**Example:** > "Your proposal-to-close rate is 23% (14/61 deals). Based on deals closed in Q3. *Gap: 12 deals missing proposal-sent timestamp — rate may be higher.*"

Actual Data Retrieval

Growth Analyst does not hallucinate. It: 1. Parses your question into intent 2. Plans which tools to call (SQL, API, vector search) 3. Executes queries against your live data 4. Synthesizes findings with citations 5. Returns answer + evidence + gaps

**Tools available:** - Lead/contact search - Pipeline/deal queries - Conversation search - Campaign/attribution queries - Revenue/analytics queries - Integration health checks - Risk/alert queries - Audit log search

Permissioned Tools & Organization Isolation

**Strict tenant isolation:** Growth Analyst only accesses data in your organization. Cross-tenant queries are impossible by design.

**Tool permissions mirror your role:** - Viewer → Read-only queries - Analyst → Read + analytical queries - Sales Rep → Own leads/conversations + team aggregates - Manager → Team data + coaching insights - Admin → All organization data

**No tool bypasses RBAC.** Every query is scoped to your membership and permissions.

Insufficient-Data Behavior

When data doesn't exist or is too sparse:

**Growth Analyst will:** - State clearly what's missing - Suggest what to instrument - Offer proxy metrics if available - Recommend manual investigation paths

**Growth Analyst will NOT:** - Invent benchmarks from "industry averages" - Extrapolate from 2 data points - Present assumptions as findings - Hide uncertainty behind confident language

Recommendations & Human Approval

**Recommendations are explicit and actionable:** - "Reduce proposal stage duration: 3 deals stuck > 21 days. Suggest adding mutual action plan template." - "Meta Ads account 'Acme_Corp' unsynced 72 hrs. Reconnect to restore attribution."

**High-risk actions require human approval:** - Bulk lead reassignment - Campaign budget changes - Automation rule modifications - Data deletion/merge operations

**Approval flow:** Analyst proposes → Human reviews → One-click execute or modify.

Limitations

| Limitation | Mitigation | |------------|------------| | Only knows what's in your data | Instrument rigorously; connect all sources | | No external market data | Provide context in your questions | | No predictive modeling (yet) | Use as diagnostic, not forecast | | Can't execute high-risk actions | Approval gate on sensitive tools | | SQL generation may hit complexity limits | Break complex questions into steps | | Vector search limited to indexed content | Ensure conversations/docs are indexed |

Next steps

  • • Connect your data sources (CRM, ads, email, chat) to enable Analyst.
  • • Start with diagnostic questions: “Where are we leaking?”
  • • Review evidence citations to trust but verify.
  • • Book a demo to try Growth Analyst on your data.