Change Order
A super points at a wall on a Tuesday and says move it. Somebody on your crew flags it — a photo, a T&M tag, a line in the daily log, a message in the project thread — and this agent turns that into a priced change order: labor, material, equipment, markup, the source it came from, and a first pass at the schedule impact. It never sends it. Every draft stops in front of whoever owns pricing.
Yours outright — no required subscription.
Every agent is built for your business — your systems, your approval chain, your way of working. We scope it on the call.
If this sounds too technical, don't worry. We take care of everything for you.
- Private VPS deployment. This agent runs on a server that belongs to you, not a shared cloud tenant. Your jobs, your pricing, your rate sheet — nobody else's data touches your box.
- A technician agent alongside it. Every hire ships as a pair: the Change Order Agent that drafts and prices, and a VPS-technician agent that lives on the same box and keeps it healthy — patched, backed up, monitored. You're not hiring one thing, you're hiring a small team of two.
- Connection to where changes actually surface. Wired into the places your team flags an extra — the field texts and photos, the daily log, the project channel, the client email thread — so nobody has to forward anything to it.
- Your numbers and your format, loaded. Loaded labor rates, unit costs, equipment, and the markup your contract allows on changes go in during setup, along with the change order format the other side requires. Drafts come out in the shape you'd have to retype them into anyway.
- Security hardening. Private VPN, firewall, and encryption configured on your server before anything goes live.
- A live walkthrough of both agents. We sit down on a call and show you exactly what the change order agent does and what the technician agent does, so nothing is a black box on day one.
- 14 days of priority support after launch, to tune it against the changes your real jobs throw at it. Then we remove our own access. You own everything — the server, the agent, the code it runs on.
Not sure it fits? Check fit in 90 seconds in the free assessment chat.
Who this is for
- You run a 15–40 person specialty contractor. Field leads write extras on paper tags, the tags ride around in a truck for two weeks, and one PM prices them at night off the estimate spreadsheet before retyping them into the format the GC wants.
- You're a remodeler or GC where the owner texts "can we move that wall" and your super says yes on the spot to keep the job moving. The change gets built that week and priced never.
- You're a 12-person agency and "one more round" landed in a chat thread in March. The month's time report shows 41 hours against a 24-hour scope line, and now you're negotiating price with a client who already has the work.
- Your change order log is a spreadsheet that's three weeks stale, so nobody can tell you the real contract value or what's actually billable on the next pay app.
How it earns trust
The failure mode of AI in change orders isn't that it gets a number wrong — people do that at 9pm too. It's that you can't tell which number came from where, so checking the draft costs as much as writing it. This agent is built around the opposite bet: every draft shows its source and its arithmetic.
Every draft carries what triggered it
The text and photo from the field, the daily log line, the answer to an RFI, the client's message — attached to the draft itself. You can see exactly what it read as a change before you agree that it was one.
The price shows its work
Line items broken to labor, material, equipment and markup, with hours times rate visible, the rate sheet it used named and dated, and the markup percentage stated. Checking a number takes about thirty seconds against the same sheet you'd have opened by hand, instead of trusting a total.
A record of what you released and what you killed
Releases are stamped — who sent it, when, which version. And every change the agent thought it saw stays visible, including the drafts you rejected, so you can tell whether it's over-reading your field chatter or under-reading it.
None of that asks you to take a number on faith — it's the same paper trail you'd want if the job ever ended up in a dispute.
Pairs well with
These share a workflow with this role. Tick any to add them to your setup.
We did the work and never got paid for it. Does this fix that?
It fixes the half you can automate. If somebody on your team flags the change — a photo, a tag, a note, a message — this agent will have a priced draft in front of you the same day instead of three weeks later. It will not fix the other half: it can't tell that a change happened if nobody says so, and it doesn't rule on whether something is in scope or extra. That's a contract judgment call and it stays yours.
It's on a T&M tag in somebody's truck. Can it work from that?
It can work from a photo of the tag the moment someone sends one — date, names, hours, material, the one-line description. It reads what's there and prices it against your rate sheet. It can't invent the hours nobody wrote down, and a tag that never leaves the truck is still a tag that never leaves the truck. The change it usually makes is that field guys start sending photos, because sending one now produces a priced document instead of a task for the PM.
The client swears they never approved it. Does this help?
It helps by making the record concrete. Every draft carries what triggered it — the actual message, photo, log entry, or answer to an RFI — and every release is stamped with who sent it, when, and which version went out. What it will not do is decide that you're covered. A friendly "yeah go ahead" in a text thread often isn't authorization, field personnel frequently can't bind the contract, and plenty of contracts waive claims not noticed in writing within days. The agent surfaces who said what and when. You decide what it's worth.
I've got a stack of change orders sitting unsigned. Will it chase them?
It will keep the log honest — what's drafted, what's out, what came back — so you can see what's outstanding without rebuilding a spreadsheet. It will not do the chasing itself. It never sends a change order out for signature, never negotiates or discounts when the other side pushes back, and never marks one approved: approval is a signature it reports, not a state it sets. It also doesn't touch money already on the books — no editing the contract value, the schedule of values, a pay app, or an invoice.
Two rounds of revisions turned into six. Does this work outside construction?
Yes. An agency scope line and a field directive are the same document problem: work happened outside what was agreed, and nobody wrote it up while it was fresh. The agent prices against your rates and the hours logged, ties the draft to the message that asked for the extra work, and hands it to whoever owns the client relationship. The vocabulary changes; the workflow doesn't.
What's covered under the $1,997 setup, and what isn't?
One company, one system of record, one approval chain. Your existing chain stays exactly as it is — the agent adds a step before a document goes out, it never removes one. Typically about seven days from kickoff to it drafting against real changes on real jobs. If you run separate pricing models per division, or several systems of record that all need to stay in sync, that's a scoped conversation on the free 15-minute call, not a price adjustment.
Hiring more than one? A department on tap — the subscription puts a build team behind every request, agent after agent.