How to audit your estimating system before you let AI touch it

Senior-estimator rule: AI should accelerate a disciplined bid desk. It should not hide messy cost codes, stale labor factors, unknown proposal formats, or review authority that only exists in someone's head.

1. Audit the data AI will inherit

Labor rates: base wage, burden, apprentice ratio, prevailing wage handling, travel rules, and overtime assumptions.
Material prices: preferred vendors, quote age, substitutions allowed, freight, tax, waste, and escalation handling.
Assemblies: which kits are trusted, which are estimator shortcuts, and which are just copied from old spreadsheets.
Historical jobs: what was estimated, what actually happened, and whether the variance was scope, labor, production, vendor, or field condition.

2. Audit the scope-control process

Before AI can help, the team needs rules for what is in the bid, what is excluded, what is alternate, and what requires a supplier or sub quote. The dangerous failures are not just wrong quantities; they are believable line items that do not belong in the scope.

3. Audit review authority

Who can approve high-dollar items?
Who can override production rates?
Who signs off on exclusions, alternates, and bid forms?
What evidence must be attached before a bid can be called ready?

4. What Vernier does with the audit

Vernier is built around company memory, review queues, evidence links, and calibrated bid state. The cleaner your estimating operating model is, the more useful the AI bid desk becomes because it can draft, compare, flag, and explain instead of guessing around undocumented rules.