Questions, answered honestly.

Probably by one of three things: the scope grew until the price stopped making sense, the person disappeared, or you ended up with something you couldn’t touch without them. I work against all three specifically. Scope and price are fixed in writing before I start. Every build ends with documentation and a handoff walkthrough. And everything runs on your accounts, so nothing is hostage to me still answering the phone.

I can’t prove I’m different in advance. What I can do is make the first step cheap enough that finding out doesn’t cost you much.

Fair question, and the honest answer is that I’m one person, so this is a real risk you’re taking. Here’s how I lower it: the system runs on your infrastructure, the code is in your repository, and the runbook — a plain-English page saying what the system does, how to turn it off, and what to check when something looks wrong — is written well enough that another engineer can pick it up. My goal at the end of a build is that you don’t need me. The retainer is a convenience, not a dependency.

If you want, I’ll put you in touch with the person I’d recommend if I weren’t available.

No. You need to know your business well enough to tell me what’s slow and what’s broken. I’ll do the translating. If a conversation with me requires you to learn what an API is, I’ve explained it badly.

Not without a person clicking approve. That’s the default in everything I build: the system can draft, sort, look things up, and prepare an action, but a person releases anything that leaves your business or can’t be undone.

Some workflows are safe to run fully unattended — moving a file, updating an internal record, sending you an alert. We’ll decide which ones together, and we start conservative.

You do, all three. I build inside your accounts using credentials you control and can revoke. The code lives in a repository you own. There’s no platform of mine you’re renting, and no license that expires.

No. Where a workflow calls an AI provider, I’ll name the provider and tell you exactly what data it sees before we build anything, and I use the settings that exclude your data from training. If a workflow would require sending something you’re not comfortable sending, we find another way or we don’t build it.

The workflow audit is $750, and it comes off the price if you go ahead with a build. A scoped build is $2,000 to $10,000, quoted as a fixed price before I start — if it takes longer than I estimated, that’s my problem, not your invoice. The retainer is $500 to $2,000 a month, month to month, cancel whenever.

There’s no long-term contract, because a contract that keeps you when you’d rather leave isn’t worth the trouble it causes.

A scoped build is one to four weeks. Most first projects land near the short end, because the first one should be small on purpose — something with a clear finish line, so you can see how I work before you decide whether to hand me anything bigger.

It will. That’s what the retainer is for, and it’s also why I document things: a change should be an afternoon, not a rebuild. If you’re not on a retainer, small changes are billed as they come, and you’re free to have someone else make them.

I don’t know yet, and I’d rather find out on a free call than sell you something that doesn’t fit. The pattern that works best is repetitive work with clear rules that a person is doing by hand — data moving between systems, requests getting sorted and routed, reports getting assembled. The pattern that works worst is judgment-heavy work where the rules live in someone’s instinct and change case by case.

If you’re in the second category, I’ll say so.