Pay-per-call advisory.

You’ve built something with AI. Maybe it’s an agent that handles a workflow you used to do by hand. Maybe it’s a Claude project you’ve been training for six months that finally sounds like you. Maybe it’s a tool stitched together with API calls and a prompt you keep tweaking. It’s mostly working. People around you are impressed. You’re shipping faster than you used to.

And there’s a knot in it.

Not a bug. A knot. The kind of thing where you can feel that something about the shape of the problem is wrong, but you can’t articulate it cleanly enough to Google it, and asking the model itself just produces confident answers in three different directions. You don’t need documentation. You don’t need a contractor. You don’t need a year of mentorship. You need an hour with somebody who’s already untied that exact knot, who can sit with you, look at what you’ve actually built, and tell you what they see.

That’s what this is.

The homepagetalks about being the human between you and every AI vendor. The Brain Lifeline is the smallest version of that promise — billed by the hour, scoped to a single conversation, no project commitment on either side. You bring the problem. I bring whatever I’ve learned from the last three years of building this stuff full-time.

The reason this offer exists is that the AI consulting market has a hole in it. On one side: $20/month subscriptions to ChatGPT and Claude. On the other side: $50,000+ enterprise engagements where someone in a suit assesses your AI maturity over a quarter. Between those two extremes is the place where most operators actually live — and there’s almost nothing built for them.

What you usually do instead is post on a Discord and hope someone with the right context happens to be reading. Or you spend three hours reading conflicting Twitter threads. Or you book a call with a vendor whose job is to sell you their tool, not to look at yours. Or you just keep tweaking, because that’s what builders do.

The cost of the knot isn’t the hour you’d spend solving it with a colleague who’s been there. The cost is the three weeks you spend solving it without one — and the architectural decision you make at the end of those three weeks, which will be the one you have to live with for the next two years.

Most of what people pay me for in these calls is the thing I almost don’t want to charge for: telling them the question is wrong. The framing they came in with — "should I use vector search or full-text" — turns out to be downstream of a decision they hadn’t made yet about how the corpus is going to be updated. The fix is upstream. An hour pays for itself the moment that gets clear.

A shared notes doc with everything we covered, written as we went. Specific next steps named by name — not "consider exploring retrieval options" but "swap to BM25 for the doc-tree side and keep vectors for the conversational side." If you want, a follow-up paragraph by email the next day with anything I thought of after we hung up. And, when it’s relevant, references to specific docs, tools, or threads that would have saved me a week if I’d known about them when I was where you are.

You also walk away with the option to keep going. Calls compound, and once we’ve done one I have context on what you’re building — the second call costs the same hour but goes deeper, faster. Some Lifeline clients book a single hour. Some book a 10-hour block ($1,250, $125/hr effective) and use it across a quarter. There’s no commitment either way.

What this isn’t: project work. I don’t write the code for you, build the brain for you, or take the work home. If we get to the end of the hour and the answer is "this needs more than a conversation," that’s when we talk about the Audit or the Install. The Lifeline holds the line at thinking together.