Why We Turned a Performance Management Tool Into an AI Studio
Vereda started as one engineering management product. It's now a studio building three AI-native products across engineering management, insurance, and real estate. Here's what changed, and what didn't.
If you've read this blog before, you know it as the place we write about async standups, 1:1 prep, and burnout detection for engineering managers. That's still accurate — Vereda AI is still a real product, and we're still building it. What's changed is that it's no longer the only thing we're building.
Vereda is now a small studio — Harrison Reid and Amy Wightman — building a portfolio of AI-native products and consulting with teams who want to apply AI seriously, without the hype. Alongside Vereda AI, we've been building PreLoss, an AI-powered home inventory app for insurance documentation, and ADU Home Resources, a resource site that helps homeowners figure out whether they can build an accessory dwelling unit. This post is about why that's one studio instead of three separate companies, and what it means for what shows up on this blog going forward.
What started as one product
Vereda AI began as a straightforward bet: engineering managers get almost no structured support, and the tools built for HR-wide performance cycles don't actually help a manager run a good 1:1 or catch a burnout signal before it becomes a resignation. That's still true, and it's still the product we write standup-bot comparisons and burnout-detection deep dives for on this blog.
Building it taught us something that turned out to matter more than the product itself: how to take a genuinely messy input — a standup response, a sentiment trend, months of scattered context about one person — and turn it into something a manager can act on in a two-minute read. That's a specific kind of AI product work, and once you've built it once, you start noticing the same shape everywhere.
Two more products, two industries with nothing to do with engineering
PreLoss applies the same instinct to home insurance: record a two-minute video walkthrough of a room, and the app turns it into an itemized inventory with replacement values and a report you can actually hand to an adjuster. ADU Home Resources applies it to real estate: instead of a homeowner spending hours parsing a zoning ordinance PDF, an AI advisor looks up the actual local rules and gives a specific answer.
Neither product started as "let's reuse the Vereda AI codebase somewhere else." They're different stacks, different data models, different failure modes. What they share is the underlying discipline of building AI that compresses a mess into one clear, trustworthy answer — and knowing exactly where AI is the wrong tool for the job, which matters just as much. We go deeper on that pattern in the next post.
Why one studio instead of three companies
The honest reason is that staying hands-on with three real products, in three industries, with three different sets of users, is what keeps our judgment sharp. It's easy to have opinions about AI product design in the abstract. It's harder — and more useful — to have shipped the thing, watched real users hit its limits, and fixed it.
That hands-on experience is also the actual basis for our consulting work. When we tell a client that a particular AI feature isn't worth building, or that a specific extraction approach will fail in a specific way, that's not a framework we read about — it's a mistake we've made ourselves, on our own products, with our own users.
What changes here, and what doesn't
The engineering-management content isn't going anywhere — standup bots, 1:1 prep, burnout detection, all of it stays, because that audience and that product are both still real. What's new is a second thread: posts about what we're actually building in PreLoss and ADU Home Resources, written the way we'd want to read them — specific, grounded in what's actually shipped, honest about what doesn't work yet.
If you're here for the engineering-management playbook, it's still all here. If you're curious what a small AI studio looks like from the inside — including the parts we got wrong — that's the new part. Either way, our projects page is the fastest way to see what we're building right now.