AI-generated reports are fast, consistent, and scalable. But they're not always right.
A market analysis might miss a local competitor only an insider would know. A team assessment might underweight a founder's track record in an adjacent industry. A risk flag might be outdated by the time a partner shares the report with their LP committee.
Until now, if something in a GemScore report needed correcting, the only option was to add a note or request a re-evaluation. The report itself was locked — a read-only artifact.
That changes today.
Edit Any Field. Keep Full History.
Admins and platform owners can now edit every field in a finalized V3 report — directly in the report view. Scores, recommendations, strengths, gaps, next steps, investment conditions, evidence chains, validation findings, even the adversarial debate.
Click "Edit Report" and every section gains an inline edit button. Make your change, optionally add a reason, save. The report updates instantly.
No re-running the AI. No waiting. No workarounds.
What You Can Edit
- Scores — Potential, Readiness, agent-level scores (Team, Market, Business, Product, Risk)
- Recommendation & Grade — Change the final call and grade band
- Executive Summary — Rewrite the TL;DR to reflect your judgment
- Strengths & Gaps — Add missed strengths, remove incorrect gaps, reorder priorities
- Next Steps & Investment Conditions — Tailor action items to the specific deal context
- Signals — Market opportunity, founder-idea fit, competitive landscape
- Complex Sections — Evidence chain, validation findings, adversarial debate (JSON editing for power users)
- Lite Reports Too — Validation verdict, top strengths, critical issues, next actions
Everything the AI writes, humans can revise.
Full Audit Trail
Every edit is logged. Who changed what, when, from what value to what value, and why.
This isn't just a nice-to-have — it's a requirement for any organization that takes compliance seriously. When a report is shared with investors, LPs, or investment committees, the edit history provides a complete chain of custody.
The report page shows a clear "Edited by Human" badge whenever a report has been modified. Anyone viewing the report — admin, partner, founder, or public — sees this indicator. Transparency is non-negotiable.
Edit history is accessible from the report toolbar. Every change is timestamped, attributed, and reversible.
Why This Matters for Enterprise
If you're running a fund, an accelerator, or a white-label evaluation platform on Athanor, you need the AI to do the heavy lifting — but you also need the final artifact to reflect your team's judgment.
For VCs and fund managers: Your investment committee doesn't want to see raw AI output. They want a report that reflects your thesis, your market view, your risk appetite. Now the AI generates the comprehensive analysis, and your team calibrates the final product.
For accelerators and incubators: Different cohorts have different contexts. A founder in Southeast Asia faces different competitive dynamics than one in Berlin. The AI provides the analytical framework; your program managers add the local intelligence.
For white-label partners: You're putting your brand on these reports. If the AI gets a market signal wrong or misses a key competitor, you need to fix it before your client sees it. Report editing means quality control happens at your layer, not ours.
For compliance teams: Regulated entities need to demonstrate that human oversight exists in AI-assisted decision-making. The audit trail provides exactly that — documented evidence of human review and correction.
AI Learns From Your Edits
Here's the part that makes this more than just a "fix the AI's mistakes" feature.
When you trigger a new evaluation on an idea that was previously edited by a human, the AI receives the full context of what was changed. It knows which fields you corrected, what the original AI values were, and what you changed them to.
The instruction to the AI is clear: treat human corrections as authoritative unless new evidence contradicts them.
This creates a feedback loop. The AI gets better at understanding your standards over time. Your edits aren't lost when the idea is re-evaluated — they're carried forward as calibration data.
How It Works
- Navigate to any completed report from the admin panel or partner dashboard
- Click "Edit Report" — the page enters edit mode with inline controls on every section
- Click the pencil icon on the field you want to change
- Make your edit, optionally note your reason, and save
- View Edit History for a complete audit log of all changes
- Restore any field to its original AI value, or restore the entire report at once
Both admin and partner panels have direct "Edit Report" buttons on the idea detail page. The report toolbar shows "Edit Report" / "Exit Edit Mode" and "Edit History" links.
What We Didn't Build
We deliberately didn't build a separate "editing interface" with a different layout. The edit happens in the same report view that everyone sees. You edit in context — seeing exactly how the change will look to the end reader.
We also didn't build approval workflows or multi-stage review pipelines. Not yet. Right now, the person with edit access makes the call. If you need tiered approvals, the audit trail gives you the data to build that process externally.
The Bigger Picture
This is part of a deliberate direction: making AI-generated artifacts useful as real business documents, not just interesting demos.
A report you can't edit is a toy. A report you can edit, with tracked changes and audit history, is a tool.
We've always believed the best startup evaluation combines AI scale with human judgment. Report editing is how we put that belief into practice — not as a philosophy, but as a feature.
Report editing is available now for all V3 reports (Screening, Assessment, and Verification tiers). Available to admins and platform partners. Get started or reach out if you're running a white-label platform and want to see it in action.
