Human First AI Consulting
Case Study

From Spreadsheet Chaos to a Job That Runs Itself

Specialty Trade Subcontractor · Commercial Construction · AI Operations Assistant

Construction runs on labor hours, and labor hours run on the honor system of whoever remembers to update the spreadsheet. We rebuilt that reconciliation as an assistant that runs itself, without asking the crew to change a single thing about how they work.

The Situation

Our client, a growing specialty subcontractor, was managing active jobs across multiple sites with crews logging time in one system, foremen posting daily production updates in a team chat channel, and a project management spreadsheet that someone had to reconcile by hand.

The owner was spending real time each day pulling reports, matching labor categories to budget line items, and trying to catch jobs drifting over budget before they became a problem. On one project, the crew had already burned roughly three-quarters of the budgeted labor allotment for a major phase of work. That number mattered. But it lived in a formula three tabs deep, and nobody had time to check it every day.

The real issue was not a lack of data. It was that the data lived in three disconnected places and demanded a person to stitch it together every single day.

The Complication

Before we could automate anything, we hit the problem that quietly kills most AI projects: the systems did not agree with each other. The category labels in the time-tracking tool did not match the category labels in the project management spreadsheet. To a person, two slightly different headings are obviously the same thing. To an automation, they are two unrelated buckets, and the whole budget calculation falls apart.

This is the part most vendors skip. We treated it as the foundation. You cannot automate a reconciliation that humans are quietly fixing in their heads.

What We Built

We designed an AI operations assistant, built on a practical and cost-conscious stack, to handle the daily grind:

1
Automated Daily Labor PullsEach day the assistant pulls the labor report from the time-tracking system and updates the project management spreadsheet by category, for every job that logged hours.
2
Production Tracking from Team ChatForemen already post daily production updates in their team chat. The assistant reads those posts and charts production against the plan, so the field team keeps working the way they already work and the office still gets clean data.
3
Budget Drift AlertsThe assistant watches each job against its budget and flags the ones trending over before the overage is locked in. A high burn rate becomes a proactive heads-up instead of a surprise at closeout.
4
A Plain-English InterfaceThe owner can ask about any job in normal language, including from the truck between sites, and get a straight answer with a strict no-guessing rule. If the assistant does not have the data, it says so rather than inventing a number.

The People-First Part

We did not ask the crews to change how they work. Foremen still post in their chat channel. Crews still clock in the way they always have. The assistant adapts to the existing habits of the field team instead of forcing new software onto people who are busy doing physical work. That is the difference between AI that gets adopted and AI that gets abandoned. The technology met the team where they already were.

The Outcome

The engagement is structured in phases, and the foundation is now in place: aligned data, automated daily updates, and early-warning budget alerts replacing manual reconciliation. What the owner gets back is the most valuable thing a small business owner has: their attention. Instead of assembling reports, they can spend that time bidding work, managing crews, and running the business.

Phase two extends the same assistant into bid management and plan handling, and phase three moves into billing and monthly draw suggestions. Each phase is scoped and priced on its own, so the client invests as the value proves out.

3
Disconnected systems reconciled into one source of truth
75%
Labor budget burn caught before it became an overage
3
Phased roadmap, scoped and priced as value proves out
The Human First Takeaway

The win here was not fancy technology. It was aligning three systems that should have agreed all along, then letting an assistant handle the daily reconciliation so a skilled operator could get back to running his company.

Client details have been kept confidential. Figures and structures reflect the scope of work delivered.

Let's Talk
← Back to all case studies