Case study: how one custom app replaced five SaaS tools for a specialty contractor

Retool's 2026 Build vs. Buy report surveyed 817 builders and buyers and found 35% of enterprises have already replaced at least one SaaS tool with a custom build, and 78% expect to build more next year. That's not a fringe story anymore. We've been doing this kind of work for smaller teams for a while, and a recent project is a good example of what it actually looks like.

What they had

The client is a 40-person specialty contractor running 150 to 200 active job sites at any time. Their stack, when we started, was five tools stitched together with email and spreadsheets:

- A CRM for sales and customer records - A scheduling and dispatch app for crews - A forms app for site inspections and punch lists - An inventory tool for materials and truck stock - A customer-facing portal for status updates and signoffs

Combined seat cost was around $4,200/month. The bigger problem wasn't price. It was that nothing talked to anything else. Their office manager spent roughly two days a week re-keying job data between systems, and crews routinely showed up to sites with the wrong materials because inventory lived in a different world than the schedule.

What we built

One web app, mobile-friendly, with five modules that share one database: customers, jobs, crews, materials, and signoffs. A job is the spine. Everything else hangs off it.

A few specifics that mattered:

- Scheduling and inventory share state. When dispatch assigns a crew to a job, the materials list auto-reserves from truck stock. If something's short, the app flags it the night before, not at 7 a.m. on site. - Forms are typed to the job. Inspections, punch lists, and signoffs are templated per job type, so the field crew only sees the fields that apply. - The customer portal is a view into the same data, not a separate product. No syncing. No "why doesn't this match the invoice." - AI-assisted development, where it helped, mostly in scaffolding CRUD, generating form schemas, and writing migration scripts. We don't ship AI features the client didn't ask for.

Build took 11 weeks from kickoff to first production users. Two more weeks to migrate everyone off the old tools.

What changed

Software spend dropped from ~$4,200/month to a smaller hosting and support figure. The office manager got most of her week back. Material-related callbacks dropped noticeably in the first quarter, though we'd want a full year before claiming a clean number on that.

The more interesting outcome: the app keeps changing. A new service line last quarter meant new job types, new forms, and a new pricing rule. With five vendors, that's a six-month project. With one app and a team that knows it, it was a week.

What stayed our job

This is the part Newsweek's follow-up on shadow IT gets right: someone has to run the thing. Hosting, backups, monitoring, security patches, and the small fixes that come up every month — that's our standing job. The client didn't hire a developer. They hired us to keep one app working so they don't have to think about it.

That's the trade. You stop paying five vendors and you start paying one team that's accountable for the whole picture. For this client, the math worked. It doesn't always. If you want to know whether it works for yours, that's what the discovery call is for.