GLF Analytics · Los Angeles

AI for Construction Management.

Your team already uses AI, scattered across the firm. We give it structure: more checks on what goes out, hours back on reports and long documents, and a more efficient operation, all inside the Microsoft 365 you already pay for.

Get started See how it works

≈5 hrs

Estimated time back per person, per week.

$1.3M

A misleading gap between two GC bids, caught in review.

8 rounds

Of team training in the first four months.

Production

A reporting workflow the firm's employees run and sign off themselves.

What you get

Time back

Ask your inbox where a project stands, what you have not gotten to yet, and who owes what, instead of digging by hand.

Better quality and accuracy

Drafts start from the full project record instead of memory, and every number is checked against the source documents.

More data analysis

Bid leveling, budget checks, and document reviews there was never time to run by hand.

One front door

Copilot is the main focus, and it runs models from Anthropic and OpenAI. If someone on your team already prompts well in another approved tool, that work carries over.

Case study

What we built at one firm.

An owner's representation and construction management firm in Los Angeles, already using AI but without structure. We organized it and set the goals: the best output for clients, with more checks and more accuracy.

The reporting workflowIn the first month, we built a reporting workflow and put it into production. It drafts each report from the project's source documents, and an employee reviews and finishes every one before it goes out.

The bid reviewWe reviewed two general contractor bids that came in $1.3M apart. The analysis showed the headline gap was misleading and caught a contingency discrepancy in one bidder's own documents.

The trained teamBy month four, the whole team was licensed and prompting on their own work. It took eight rounds of training, customized to each person and their role.

Estimated hours back, per person, per week
Where the time comes back	Est. hrs / wk
Project status reporting	1.5
Email and project-history review	1.5
Meeting follow-up	1.0
Document checks and bid review	1.0
Per person, per week	≈5 hrs

≈50 hrs

per week across a team of ten project managers

≈2,600 hrs

per year, back in the business

All figures on this panel are estimates based on the workflows deployed, not measured client data.

One week there was nothing significant to report, and the system said exactly that. I'd rather have that than three paragraphs of filler.

The names stay private. The longer arc is on the track record page.

The trial

Private workshop.

$250 per person

A private session for you, or a small group of your decision-makers. We work the five workflows below on your own projects, and you leave with every prompt we used.

Own your inboxRebuild a project's history from email and surface what needs action. Outlook + Copilot
The Report PipelineProject source material into a structured Word report, ready for your review. Copilot + Word
Meetings into actionDecisions, owners, and follow-ups pulled out of every meeting. Copilot + Teams
Compare two bidsTwo contractor bids leveled, beyond the headline number. Copilot + Excel
Email into dataYears of email turned into a structured table you can analyze. Copilot + Excel

Set up a session

Full engagement

90-day AI enablement.

Reach out for pricing several options

Every firm runs differently, and so does each person in it. We take your company from scattered AI use to a working system, customized for each employee.

Weeks 1-4 Assess and plan.We survey your employees and sit with the work. You get a roadmap: where AI helps your firm, where it does not, and in what order to build.
Weeks 5-8 Implement the roadmap.We sit with each employee on their own work: which tools fit, prompts proven to work, and output that sounds like you and not like AI.
Weeks 9-12 Training and custom builds.Training for the whole team, plus the custom pieces your firm needs: templates, document generators, whatever makes it stick.

What you own after: working workflows, a prompt library organized by role, and a team trained to run the upkeep itself, from updating prompts to catching where the output misses the mark.

After the 90 days: support as needed. No permanent retainer.

See what the 90-day engagement looks like in detail

Book the 30-minute call

FAQ

Questions we get.

Is this going to replace people on my team?

No. It gives your existing people a state-of-the-art assistant, and they still review and sign off on everything before it goes out.

My team is already experimenting with ChatGPT on their own. Is that a problem?

It's the normal starting point. The work here turns that experimenting into shared workflows with rules: which tools are approved, what data goes where, and who signs off.

Where does our project data go?

It stays in your Microsoft environment, inside your own tenant, and client data never runs through consumer AI tools. Privacy is the first question we ask about any tool, and if it can't answer it cleanly, we don't recommend it.

What tools do you work with?

Copilot, Outlook, Excel, Word, SharePoint, and the automation layer that connects them, because that's where construction management firms already live. We also work in and teach Claude, Claude Code, and ChatGPT, and when someone on your team already prompts well in one, we build on it under the same data rules.

How much AI ends up in the final work product?

As much as earns its place and no more; every workflow keeps a named reviewer and the final words are always yours. Clients can tell when a report reads machine-written and nobody checked it, and that costs trust.

Get started.

Two ways in: set up a workshop for your decision-makers, or book a call and we talk about your firm.

Every note comes straight to Gabriel Freeman, GLF's founder, who runs every engagement himself.

Set up a session Book a 30-minute call Or email us a question