Connect your ServiceTitan or Housecall Pro once. Then just ask: find the profit leaks, design the incentive plan, run the pool, pay the techs. Every answer is benchmarked against hundreds of trades companies.
npx skills add sharewillow/engine
·
Simple credits, cancel anytime
ServiceTitan and Housecall Pro don't agree on what a job, an invoice, or a callback even is. We've spent years normalizing every record and encoding the rules, so when Claude answers, the number is right.
Ask a generic AI on raw APIs and it will confidently get it wrong. Watch both answers come back. Same shop, same data.
Hundreds of edge cases like these, already handled: encoded rules, normalized data, cross-company benchmarks. Without the ShareWillow MCP, the answers aren't just incomplete. They're confidently wrong.
Plenty of tools have one of these. The Engine is the only place all four live behind a single connection in Claude.
Clean, normalized reads of ServiceTitan and Housecall Pro (jobs, revenue, callbacks, memberships, close rates) with years of edge cases already handled. Clean enough to calculate a bonus from, because that's exactly what we do with it.
Every comp design pattern we've proven across hundreds of shops (SPIFFs, tech scorecards, callback pools, membership incentives) installed into Claude with one command, free and open source.
How your shop compares to hundreds of trades companies and, uniquely, which plan designs actually moved each metric. Not survey data. Real plans, real payouts, real outcomes.
Design is a conversation. Execution is ShareWillow: pools calculated every period, distributions prepped for your one-click approval, techs paid on time. The plan doesn't live in a PDF. It runs.
These aren't prompt templates. Each master skill is years of performance-pay operations encoded on the ShareWillow MCP: normalized field-service data, cross-company benchmarks, payout rails. You could spend months rebuilding the logic from scratch. You still couldn't rebuild the data.
All skills are free and open source, and every one can be scheduled to run on its own: weekly anomaly watches, month-end pool runs, quarterly benchmark refreshes. They come alive connected to the ShareWillow MCP.
Everything below is a real workflow the Engine runs today, in Claude Desktop or Claude Code, no terminal required.
Anyone can pull your numbers. Only ShareWillow knows what hundreds of trades companies did about theirs: which plan designs they ran, what they paid, and what happened to callbacks, tickets, and memberships afterward. Every payout we've processed is a labeled outcome in the corpus your answers draw from.
repair sales per tech after launching their plan, with $156.8K paid to techs.
memberships sold per tech (0.2 → 3.8 per month).
average ticket ($80 → $284) after tying performance pay to it.
revenue per tech; company revenue up 91% on 10 techs.
tech-generated leads, with $146.2K paid out on the plan.
payout growth in 5 months ($2,890 → $100,329/mo) across 7 plans.
When the Engine says "shops like yours pay $25 per membership and it moves the number 2–4×," that's not a model's opinion. It's the median of shops that actually did it. Always anonymized: no shop, tech, or customer is ever identifiable.
"I would 100% encourage people to do performance pay, and definitely partner with ShareWillow."
Chris McDowell · Business Operations Manager, Clog Busters (31 techs · $1.01M in bonuses paid)
And unlike a consultant, it doesn't leave after the invoice. It runs the plan every payroll period.
And you stay in control the whole way: the Engine drafts, you publish, and every dollar requires your sign-off.
The skills are free and open source. Credits meter what's valuable: your live data and our benchmarks. Unused credits roll over; they never expire.
Already a ShareWillow customer? Unlimited credits are included in your subscription, plus the one thing no prompt can do without us: actually paying your techs.
No. The Engine works in Claude Desktop, the same app you'd use to chat. Connect your field-service software once, then talk to it like you'd talk to your best ops manager. Claude Code works too if that's your thing.
Benchmarks are always anonymized and aggregated: no shop, tech, or customer is ever identifiable, and small groups are never reported. Your raw data is yours; it's never shared or sold.
No. The Engine drafts; you publish. Every plan goes live only when you approve it in ShareWillow, and every distribution requires your explicit sign-off before a dollar moves.
A long way. A full audit is ~150 credits, a plan design with simulation ~50, and day-to-day questions cost 1–5 each. Most single shops use 800–1,500 credits a month, and unused credits roll over, so heavy months are covered.
Jobber and QuickBooks are next, with more platforms to follow. You can also start in demo-data mode or with CSV exports to see what the Engine finds.
A consultant gives you a plan in six weeks for $10–25k, based on experience with a handful of shops. The Engine audits your real data in minutes, benchmarks against hundreds of companies and 14,847 payouts, then runs the plan every payroll period, forever.
We're onboarding a small group of shops, consultants, and roll-up operators this quarter. Grab a spot.
Get on waitlistor install the free skills: npx skills add sharewillow/engine