Lyonsmith TechnologyAI Automation

Client work · Case study

The Bid Tool: 4 hours of bid work, done in about 10 minutes.

Built for a residential home builder whose estimator compared every subcontractor bid and filled out the cost sheet by hand. Details below are process-level and anonymized. The numbers are real.

Client  A residential home builder Category  Bid comparison & data entry Status  In production

The problem

Every home started with four hours of bid comparing, scoping, and data entry.

Subcontractor bids arrive for every trade: PDFs, Word docs, sometimes a photographed handwritten page. For each house, the estimator read them all, lined up what each sub actually included, compared prices line by line, and then typed the winning numbers into the master cost sheet by hand. Skilled estimating time, spent on reading, cross-checking, and typing. And it scaled the worst way possible: more houses meant more hours, or more payroll.

What was built

Bids in. A side-by-side comparison. Excel filled in a minute.

The Bid Tool ingests every subcontractor bid for a trade and lays them out in a comparison dashboard, line item by line item, so the estimator can weigh scope and price in one view. The estimator picks the winners; the tool never chooses. Once the selections are made, the tool moves them into the company’s existing Excel master cost sheet: about 1,800 cells populated in under a minute. Anything the tool is not confident about gets flagged for review instead of guessed.

01

Bids in

Subcontractor bids arrive per trade: PDFs, Word docs, even photographed handwritten pages.

02

Compared side by side

Every bid for the trade lands in one dashboard: one row per line item, one column per sub, so scope gaps and price differences are obvious at a glance.

03

The estimator picks

The human selects the winning bid for each trade. The tool never chooses, and anything it is unsure of is flagged for review.

04

Excel filled in

The selections populate the existing master cost sheet: about 1,800 cells in under a minute, in the format the team already used.

The results

The centerpiece: the before and after.

Before
~4 hours
per house
Reading, comparing, and retyping subcontractor bids

Every house tied up skilled estimating time in bid comparing, scope checking, and data entry before the real work, pricing judgment, could even start.

After
~10 minutes
per house
Compare on one screen, pick the winners, done

The estimator scans the comparison dashboard, makes the calls, and the cost sheet fills itself: about 1,800 cells populated in roughly a minute.

~950 hrs/yr
given back, roughly half a full-time position

Three hours and fifty minutes back on every house. That’s not time saved in the abstract: those are hours the company was paying for, every year, now spent on work that actually needs a person.

Watch it work Navigating the tool, start to finish
Walkthrough video coming soon
The problem, explained Why four hours per house, in one minute
Explainer video coming soon

Where this engine shows up next

The same document-processing engine drives two portfolio builds.

The Document Intake Desk applies it to invoices, receipts, and contracts. Shortlist applies it to resumes. Both run live, right now.