AI Dispatch Won't Fix a Bad Dispatch Strategy

AI dispatch is real, and it works. Routing technicians by location, drive time, skill match, parts on hand, and job priority — and re-optimizing the day as cancellations and overruns hit — reliably delivers 12–20% more jobs per tech per day and 15–25% less drive time. Some platforms report 25–40% gains in daily job completion. The technology is no longer the question.
The question nobody asks before they buy: what are you optimizing for? Because AI dispatch will faithfully maximize whatever objective you hand it — and most contractors never define one. They turn on the routing module, it crushes drive time, and six months later margin is flat or worse. The engine did its job. The strategy was missing.
If this is you
You turned on AI routing. Drive time's down, the techs are completing more calls, the board looks gorgeous. And the P&L is no better than last year. Here's why: the system optimized for "more jobs," so it filled the day with the easy, low-margin work and quietly de-prioritized the big, messy, profitable jobs that don't slot neatly into a route.
What AI dispatch actually does well
Give it credit — this is genuine productivity:
- Evaluates every open job against every available tech simultaneously, factoring GPS location, drive time, certifications (EPA 608, trade license), truck inventory, and customer tier — then recommends the best assignment.
- Re-optimizes dynamically when a job runs long, a customer cancels, or an emergency lands, instead of letting a static morning route fall apart by 10am.
- Cuts the unbillable windshield time that quietly eats 15–25% of a tech's day.
- Learns from dispatcher overrides so the recommendations get sharper over time.
For a labor-short contractor, that's exactly the kind of leverage you want — more completed work out of the same crew. (It's one pillar of the operating layer for the trades.)
Where it goes wrong: the objective function
Here's the part the FSM demo skips. An optimizer is only as good as the thing you tell it to maximize. The default objective in most dispatch tools is some version of "jobs completed" or "minimize drive time." Both are proxies for productivity — and both can quietly destroy margin:
- Optimize for most jobs → the system favors short, easy calls and buries the high-margin replacement work that takes longer to slot. You complete more tickets and make less money.
- Optimize for least drive time → the system clusters geographically and may route your senior tech to whatever's nearby, not whatever needs his skill. Cheap to get there, expensive in opportunity cost.
- Optimize for SLA / response time → great for contracts with penalties, but it can starve the discretionary work where your real margin lives.
None of these are the tool's fault. They're the absence of a decision. The right objective for almost every contractor is gross profit per technician hour — and almost no off-the-shelf dispatch engine optimizes for that out of the box, because it can't see your margin. Margin lives in the seam between the FSM and the books, which is exactly the data your field software doesn't show you.
Free cash flow audit
Books behind? We rebuild from the bank statement up.
We benchmark your books against 2,200+ service businesses and tell you exactly where the money is going.
Strategy first: decide what you're maximizing
Before you let an algorithm run your board, make four calls a human has to make:
- Define the objective. Gross profit per tech hour, not raw job count. If the system can't weight jobs by expected margin, you have to feed it a proxy — job type, ticket size, segment — that correlates with profit.
- Protect the high-margin work. Make sure replacement, diagnostic, and high-ticket jobs get prioritized, not de-prioritized for being "inefficient" to route.
- Match skill to margin, not just to proximity. The closest tech isn't always the right tech. Keep your scarce senior hours on the work that needs them.
- Decide where the human overrides. Big jobs, key accounts, judgment calls — the dispatcher (or owner) owns those. The AI runs the routine. (This is the same logic as the AI operating system for your money — automate the reflex, keep the judgment human.)
Get those four right and AI dispatch becomes a margin multiplier. Skip them and it becomes a very efficient way to do more of the wrong work.
The honest take
I'm not anti-software — the productivity gains are real and you should capture them. I'm anti-buying the engine and skipping the strategy, which is what I see constantly. The vendors sell the routing horsepower because that's what demos well. Nobody sells you the objective function, because that requires knowing your margins — and that's your job, or your finance layer's, not the FSM's.
Dispatch is the control center for field profitability. Treat it that way. The software is necessary; it is not sufficient. Decide what you're optimizing for first — then let the AI run flat out.
FAQ
Does AI dispatch software actually work?
Yes. It reliably delivers 12–20% more jobs per technician per day and 15–25% less drive time by routing on location, skill, parts, and priority and re-optimizing in real time. The productivity gain is well documented across platforms. The caveat is that it optimizes whatever objective you set — so the strategy around it determines whether the gains hit your margin or just your job count.
What should AI dispatch optimize for?
Gross profit per technician hour, not raw job count or pure drive-time minimization. Maximizing jobs tends to favor short, low-margin calls; minimizing drive time can route senior techs to nearby low-value work. Since most dispatch tools can't see your margin, feed them a proxy (job type, ticket size, segment) that tracks profitability.
Do I need to replace my FSM to get AI dispatch?
Usually not. ServiceTitan, Jobber, Housecall Pro, and FieldEdge have routing modules, and standalone AI dispatch can layer on top. The bigger gap is the profit logic, which lives between the FSM and your accounting system — that's the operating layer, not the FSM.
Why did our margins not improve after we turned on AI routing?
Almost always because the system was optimizing for job count or drive time, not profit. It filled the day with efficient, low-margin work and de-prioritized the big profitable jobs. Reset the objective toward gross profit per tech hour and protect the high-margin work from being routed away.
If you've got AI dispatch running but can't say whether it's optimizing for profit or just for a tidy board, that's the strategy gap — and it's costing you margin at full efficiency. Book a 15-minute call and we'll map your dispatch objective to gross profit per tech hour. The first profitability audit is free.
Get the next one
Want next week's benchmark in your inbox?
One email a week. Real numbers from 2,200+ service businesses. No fluff. Unsubscribe anytime.
Related reads
Operations
AI for Contractors: The Operating Layer for the Trades AI Can't Replace
AI won't replace your best tech — there aren't enough of them. It replaces the management layer around the crew. What AI actually runs in field operations, and what it's worth at a $5M contractor.
Operations
First-Time Fix Rate: The $52K/Month Callback Leak AI Actually Fixes
Industry first-time fix rate sits at 77%. Every callback costs ~$650 and bills nothing. Why 4 of 5 callbacks are information problems, not skill problems — and the gross profit per tech hour it's quietly draining.
Operations
You Can't Hire Your Way Out: Revenue Per Technician in a Labor Shortage
HVAC is short ~110K techs and plumbing faces a 550K shortfall. You can't hire your way to the next tier. The only lever left is revenue — and gross profit — per technician. Here's what good looks like and how to move it.

About the author
Sam Young
Founder & CEO
Founder of Level — the AI operating layer for contractors and skilled trades, and the other operating businesses where scarce labor is the constraint. Ex-CFO across trades, SaaS, and service businesses. 4 years as Director of Growth Product at BuildOps, building financial tooling used by 1,000+ commercial contractors. Four years in PE and investment banking rolling up and acquiring service businesses — $2.5B in total transactions including M&A and IPOs. Stanford MBA, Brown undergrad. Level operates its own proprietary benchmark research (2,200+ companies, $13.25B in revenue analyzed) which informs every client engagement.
LinkedIn