Make every team operate
like your best one
SpireField AI reads your ServiceTitan, payroll, accounting, and expense-card data. It tells every team what to do today to increase revenue, and takes the actions with your approval. More first-time fixes. More replacement wins. Higher profit per truck roll.
Actions today
Owner / exec view🧠 It runs on your best people’s judgment, and it gets sharper. Learn more →
For managers · Actions Today
Every morning: taking action to increase revenue
Your field teams follow different standards. Actions Today runs every team like your best manager would.
Instead of spending the day reacting to whatever comes in, managers start it playing offense. Targeted actions for specific jobs, techs, and dollars to grow the business, ready before the first truck rolls.
Prep replacement quotes for Stubbs and Zamora, before the tech knocks
Escalate Cliff Villareal, 3 visits in 17 days, to a replacement conversation
Pair Nick Foss with a senior tech (5 callbacks this month)
Honest about gaps: No warranty claim data loaded. Can't check warranty recovery rate · No fleet/GPS data. Fuel cost per job not calculable
claude-sonnet-5 · $0.22 per run · replay of a real run on sample data
Ask a follow-up. The AI keeps the full context
For techs · Job Predict
Every job, pre-screened before the truck rolls
Your technicians go into homes without guidance. Job Predict tells them the fix, parts, price, and replacements, before the truck rolls.
New techs perform like veterans sooner, and techs who succeed stay longer.
Job Predict: Carrier Central AC
tomorrow’s boardCustomer reports: "AC blowing warm again." 2011 unit, third service call this season.
Likely diagnosis
Parts to have on hand
Price range
$349 – $890
from your own price book
Replacement flag
⚠ 3rd repair this season on a 2011 unit, about $1,900 spent. Take the $9,800 replacement quote to the door
🧠 Encoded rule applied to this job
Any customer with 3+ service calls in 30 days on the same equipment should be flagged for a replacement conversation, not another repair.
Service manager weekly huddle, May 2024 · cited on 8 jobs this month
claude-sonnet-5 · $0.10 per run · replay of a real run on sample data
For everyone · Ask Anything
Access and use your data with a simple chat
Your data is hidden across complex dashboards. Ask Anything gives you the answers to get more profit.
This makes it easy to use and grow. Across ServiceTitan and every major system you run, no training required.
ServiceTitan + ADP payroll + Sage accounting + Ramp cards
Sample questions. Tap one to see it answered
If your team can send a text, they can use this. Use right away, not after months of onboarding. ServiceTitan stays your system of record. This is your system of intelligence. No report builder, no new screens for your team to learn.
Every action is specific, sourced,
and priced in dollars
You approve.
The AI does the work.
SpireField AI surfaces specific actions to execute: the exact job, the exact tech, the exact part number, the dollar value, and who owns it. You approve, and the AI carries it out.
Drafted parts order: tomorrow’s board
Drafted by the AICascade Supply Co. · Order desk, Branch 12
Account #4471 · Will-call pickup under “Summit, Mike R.”
Run capacitor 45/5 µF 440V ×2
median of 39 of your invoices
$80.00 ea
Contactor, 2-pole 30A ×1
median of 78 of your invoices
$85.00
Condenser fan motor, 1/4 HP ×1
no invoice match found
TBD, desk will quote
Subtotal
$245.00 + 1 line to be quoted
Drafted by the AI. Sent by software. Approved by you, every time.
Send the parts order
Drafted PO emailed to the supply house: part numbers, quantities, account number, will-call pickup. The Approve demo above is the real flow.
The right part on the truck = the job done in one visit, not two.
Follow up on every stale estimate
The AI finds the quotes going cold and drafts a personal follow-up to each customer: their specific quote, their equipment, the season. Approve the batch → sent.
Quoted work that never closed is the cheapest revenue there is.
Renew every expiring membership
Drafted renewal notes built from each customer’s own history, plus winback offers to customers with aging equipment you haven’t seen in 18+ months.
At the sample company: a $56K/year membership gap between best and lagging brand.
Keep every truck stocked and tracked
Stock levels computed from what each tech actually uses. Restock orders drafted weekly → approve → sent, and truck inventory stays current. No scanning, no counting.
A pain every shop knows, solved from the paperwork you already generate.
Flag the dispatch board
“Reassign job 9914 from Nick Foss to Tom Brandt: double-booked, and Foss is at 5 callbacks this month to Brandt’s zero.” One tap applies it in ServiceTitan.
At the sample company: one tech’s callbacks were ~$665/month in wasted labor.
Replacement quote ready before the tech knocks
Draft estimate written onto the job in ServiceTitan, built from the customer’s own repair history. What they’ve already spent does the selling.
At the sample company: the average replacement deal is $9,424.
Plus: 5:30 AM pre-job briefings for every tech, warranty claims on parts that failed in-warranty, coaching packets for high-callback techs, context-aware confirmation calls. Every one behind the same approve button.
You approve the policy once, “yes, always order tomorrow’s parts the night before,” and the AI executes every instance.
What one overnight run found
Every number below came from the examples you just saw. One pass over the sample company’s data.
$9.4K–$18.8K
of replacement opportunity on tomorrow’s board, flagged in time to prep the quotes
$25,300/mo
replacement close-rate gap between the best and middle brand
$18,700/mo
gross-margin gap between the best and lagging brand, with the two causes named
$56K/yr
membership-revenue gap between the best and lagging brand
The same overnight pass runs on your numbers in week one.
Messy data? Missing fields?
That’s normal, and it’s fine.
Most teams don’t fill in every field, and this tool was built for exactly that.
No required fields
It reads whatever columns exist and works from what’s there. Nothing breaks because something wasn’t entered.
Runs on the records you already create
Every dispatch and every invoice generates data automatically, and most of the answers above run on those records alone.
Honest when something’s missing
When the data can’t answer something, it says “couldn’t check, not in the data” instead of guessing. You saw it on the actions list above.
It runs on your best people’s judgment, and it gets sharper
Your strongest technicians and service managers know things that aren’t written down. SpireField AI captures that from tech trainings, ride-along recordings, meetings, and expertise you select. It then encodes that expertise into the AI agent and applies it to every job, at every brand. You saw one on the Job Predict card above.
1 · Captured
“Every time Lennox tech support says board, I check the inducer first. Eight out of ten times that’s all it is.”
🎙 Ride-along with Mike Chen (senior tech), June 2024
2 · Approved by you
The quote becomes a proposed rule. You accept, edit, or reject. Nothing changes behavior silently.
✓ Accepted3 · Applied & cited
The $312 inducer ranked first, the OEM’s $1,480 board last, because of this rule. Cited on 12 jobs this month.
And it grades itself: every prediction is checked against the actual close-out in your next export. No one clicks right or wrong. Misses become proposed rules. The system gets sharper from its own mistakes.
Walk through the product →Same standard, every brand
You bought ten companies. Now you can see which ones are leaving money on the table, and exactly where.
Basin runs a 45.7% gross margin to Ridgeline’s 60.5% on the same work.
The gap is parts cost (29% higher per job) and OT-driven labor. Closing it is worth ~$18,700/month in gross profit.
Ridgeline closes replacement leads at 35.7%. Clearwater: 26.5%.
At Clearwater’s lead volume that’s ~$25,300/month left on the table. A close-process gap, not a lead-volume gap.
Same playbook rules, applied at every brand. Same data, read the same way. The gaps stop being anecdotal and start being specific, with a number and an owner attached. And as your teams see this drive their daily work, the data gets better on its own, because it’s the first time filling in a close-out field comes back as something useful the next morning.
See it on your own numbers
Runs on the exports your systems already produce. One saved ServiceTitan report, dropped in. Live in week one, no integration.
Get in touchThe demos above replay real product runs on a fictional company’s sample data.