Calibration
Review the AI's interpretation and finalize your first report into a calibrated baseline.
Calibration is the step between a fast first draft and a baseline you can stand behind. It lets you review the AI's interpretation, correct missing context, and finalize the report as a calibrated v2.0. Available on Pro and Enterprise plans.
Why Calibrate?
When Compass generates a report, the AI makes informed assumptions based on your connector data and industry norms. But every organisation is different:
- Your "high" orphan account rate might be expected due to a recent merger
- A low MFA adoption rate might be intentional for shared service accounts
- Provisioning times might include a manual approval step that's by design
Calibration lets you tell the AI: "Here's what you understood correctly, here's what needs to change, and here's the context that should shape the baseline."
How It Works
- Open Calibration — From your report, click Review Before Finalizing
- Review the analysis plan — Compass shows how it understood your problem, which data it used, and which assumptions guided the report
- Correct what's wrong — You can confirm or adjust the problem framing, approach, data sources, and targeted clarification questions
- Finalize the baseline — If everything already looks right, Compass promotes the report directly to a calibrated
v2.0. If you made changes, it regenerates the affected sections before saving that calibratedv2.0
Calibration is non-destructive — your original report version is preserved, and completing calibration creates a calibrated v2.0 report. Calibration is currently a one-time step per discovery; later updates should use reruns.
What You Can Calibrate
- Problem framing — "The real priority is access certification debt, not provisioning speed"
- Analysis approach — "Use a stricter lens because this is a regulated environment"
- Data coverage — "Include this connector" or "exclude this source from the baseline"
- Operating context — "This pattern is expected because we're mid-migration" or "this control is intentionally manual"
Best Practices
- Calibrate after your first report to establish a baseline that matches your environment
- Focus on the top 3–5 assumptions that have the biggest impact on your findings
- Use reruns after major organisational changes (mergers, migrations, policy updates) to measure what changed against your calibrated baseline
- Use calibration notes as documentation for auditors who need to understand your IAM decisions