Breasy is an exterior-maintenance marketplace connecting property managers with the local providers who do the work. We served 250+ providers completing 10,000+ jobs a year across 10 markets, and we did it on a custom-built platform instead of a stack of off-the-shelf tools.
When we raised, the pitch was "we do the thing." The proof had to be that we were actually doing it, which meant building the thing before anyone else would.
The thesis
Service providers lose hours every week to tools that weren't built for the way they actually work. Quotes scattered across email, jobs tracked in notes apps, pricing estimated on gut. The tax on being a small service business is coordination overhead, and the tools that claim to fix it are either overbuilt for enterprise or underbuilt for anything past a solo operator.
We built it inside-out
Here's the key thing we did differently. We didn't start by building an app for customers. We started by building for our own fulfillment team.
The first job of the platform was to make our internal coordination faster: to let the fulfillment team triage their workflow, move a job from request to done without dropping a ball, and stop bleeding time to manual handoffs. We made our own operation fast and reliable first, on software we controlled and could change the moment something wasn't working.
Only then did we turn it outward. Once the efficiency was real for us, we built the customer-facing app that handed the same gains to the SMBs on the platform. We weren't selling them a promise. We were giving them a workflow we had already proven on ourselves.
How we built it
As the product and technical lead, I directed the build end to end with a small team: a designer and three developers. The architecture calls were mine. The ones that mattered:
- AWS and Bubble.io as the backbone. Bubble for the speed of iteration a seed-stage team needs, AWS for the primitives (storage, functions, queues) that Bubble couldn't handle cleanly. Everything that changed often lived in Bubble, everything that had to be reliable lived in AWS.
- Web and mobile from day one. Providers live on their phones during the day and at desks at night. The platform had to work in both without feeling like a port of the other.
- AI pricing models. A quote-assessment engine that read incoming jobs and returned a recommended bundle and price range, trained on our closed-deal data. Not automation for its own sake, but removing the "am I underpricing this?" anxiety that slows every service operator down.
- An automation layer (n8n and a custom Chrome extension). The work that didn't need human judgment, like status updates, follow-ups, and internal handoffs, got offloaded to workflows. Coordination time across the team dropped 60% end to end.
- Quote-to-close tooling. Tightened the sales motion and lifted close rate from 50% to 70%.
What changed
The platform ran, live and at scale, and the scale it supported was built by a team that stayed lean the whole way. No 30-person engineering org, no agency build, no enterprise-stack rebuild. Five people with taste, opinions, and the right tools for the job.
This is the proof point I reach for when a small-business owner asks "can I really run a real platform without spending $500K?" The answer is yes, if you pick your stack like you mean it, and if you build it for the people doing the work before you build it for show.
By the numbers
- Built inside-out: our fulfillment workflow first, then the customer app
- 250+ service providers, 10,000+ jobs/year across 10 markets
- Coordination time −60% via automation
- Close rate 50% → 70%, revenue $1M → $2M ARR
