What We Didn't Build

Last winter, my wife was running a senior-level search the way she runs everything else: with a system, and to a standard. I watched the tools fail to keep up with her.

We went looking. Everything we found was built for somebody else. Resume builders for early-career applicants. Trackers that recorded what you'd done without improving what you did next. Text generators with no sense of what makes a senior candidacy compelling. Expensive career coaches who still left the weight on your desk.

People at that level already know how to do this well. What they lack is infrastructure that holds the standard across months, while the rest of life keeps its own schedule.

So I built what was missing. At first it was just a way to stay organized, put together late on a Tuesday night. It worked far better than either of us expected, and then it stopped being a personal workaround and started being a product.

That product is Exec Foundry, and it's live at execfoundry.com. Catalyst Forge operates it.

What it does

You put your resume and background in once. You name a direction, or paste in a role you're considering. A few minutes later you have materials, interview prep, and a sourced company brief for that pursuit, all off the same record.

A serious search can run for months. The workspace is built for the duration. Each pursuit keeps its own story, its people, and where things stand, so when a recruiter calls back in October about a conversation from July, you open one page and you're caught up. That's the product: the workspace that keeps the twentieth pursuit at the standard of the first.

What we didn't build

There's an obvious feature we turned down: auto-apply. Point it at a job board, let it fire applications on your behalf, watch the number climb. Plenty of products in this category do exactly that. It demos beautifully, and it's the first thing people ask for.

We said no, and we're not planning to change our minds.

In every search we watched, volume worked against quality. The person on their fortieth application sends weaker material than they sent on their first, not through laziness but because nobody holds that standard alone for that long. We wanted the fortieth to be as good as the first, not to maximize how many went out. A tool that fires applications automatically solves the wrong problem with more horsepower.

Positioning is the work. Documents are the output. We automated the preparation, not the decision to apply, because choosing which story matters for this role, or which gap to name, is judgment we weren't willing to hand to an algorithm. So the product carries the hours and gives the thinking back. You review the prep. You shape the narrative. You decide when it's ready. The product stops at the part that makes you worth hiring.

That's the same thing I tell consulting clients. The goal was never to hand your work to a machine. It's to stop trading your best thinking for volume so you can spend it where it counts.

Three things shipping it taught me

Generic output was worth nothing here. For a senior candidacy, the gap between a free resume generator and a product people pay for monthly came down to grounding: it starts from your actual history and the way you describe your own work. Without that, the AI writes confident, plausible text that no hiring manager needs to read. The hours I spent on grounding paid for themselves. Clever prompting, without that grounding, mostly did not.

The unglamorous engineering is most of the job. Early on, a cover letter went out with the user's login handle where their name should have been: a small, ugly bug, and exactly the kind a demo never surfaces. Fixing it properly meant reworking how identity flows through document generation, and most of the real work has looked like that. Getting the AI to write something was never the hard part.

Getting to reliable is a process problem. Exec Foundry took over two thousand commits, and I was in it every day. A lot of that time went into learning how to build this way, and that part came out as ForgeTrail: the phase, the decisions, and the gotchas live in the repo, so the next session starts where the last one stopped. With that in place, a small app on foundations we've already settled takes a day, sometimes two. Start from a blank session and a weekend gets you a demo, because the AI writes the feature and then you re-fight every bug the last project already settled.

A day-old app and a product are different claims. Exec Foundry generates documents for paying users and has to keep doing it while I'm asleep. That reliability came from months of hardening after the first working version, and no amount of speed on the first build substitutes for it.

Where it stands

Exec Foundry is live, it's paid, people are running searches on it, and I still harden it every week. There's a free trial with no card if you want to look around, and my wife tells her side of the story at Our Story.

I didn't set out to build a product. I set out to solve one person's problem, and it turned out a lot of people had the same one.

I built it so the search could meet the same standard as the career behind it.