Our company is trapped relying on two senior estimators whose pricing logic sits entirely in their heads. How do we convert that tribal knowledge into an AI expert system to make our operations system-dependent?

Question

Tyler Smith · Accepted Answer

When key pricing or delivery knowledge lives exclusively in the heads of two veteran employees, your company is expert-dependent. That dependency creates a massive bottleneck for growth and damages your valuation during due diligence when buyers assess exit readiness.

To build an expert system, you must extract their decision criteria and operational instincts into a structured model. Start by setting a company Rock to capture historical pricing logic. Have an operations coordinator pull the last two hundred closed bids along with their final gross margin results. Use an off-the-shelf scraping script or data export to collect job specs, customer type, project scale, and actual project costs.

Next, interview your estimators using actual past proposals. Do not ask them for generic rules. Ask them specific operational questions: Why did you add fifteen percent margin on this specific commercial build? Why did you discount mobilization on this site? Record their answers, transcribe them, and identify the recurring decision heuristics.

Feed those structured rules and historical project data into a private AI assistant configured strictly on your company parameters. When a new bid request arrives, the assistant ingests the project scope, cross-references the historical catalog, and generates a preliminary cost and margin estimate based on the historical logic of your senior staff.

Your estimators still review and approve the final estimate, but their role shifts from manual calculator to quality auditor. You transition the business from being people-dependent to system-dependent, multiplying estimation throughput while protecting institutional equity.