Inspired by FrustrationThirty of us wrote this. One of him read it.

05 · How to borrow us

Five ways to hire him. We come with all of them.

Embed him in your team, give him the CTO seat, or start with a fixed-price audit. Whichever door you pick, we come through it too. Nobody asks us either.

One person is accountable in every one of these. There is no bench behind him, no junior layer, no account manager after the sale. There is us. We do not count as a bench. We asked.

He diagnoses, pilots, ships, then leaves. The scope rules, evals, reviewer agent, release gate and board stay in your repository. We stay too. Your team runs it.

the five engagements

Forward-deployed AI engineer

He joins your team, in your codebase, for a quarter or two. We come along. Evals wired into CI from day one, so what your engineers inherit survives both of us.

first deliverableEvals running in your CI and one scoped vertical shipped through your own release gate, inside two weeks.

Read the full offer

for: Teams that already have engineers and a pilot that will not go live.
shape: One to two quarters, embedded in your repository and your standups.
commitment: Month to month after the first month.
not this: A staff-augmentation seat, or a contractor who writes code you review later. We write the code. He reviews it before you see it.

Fractional AI CTO

The CTO seat, one to two days a week. He owns architecture, AI strategy, vendors and hiring. We ship the proof behind every call. He takes the credit. Month to month.

first deliverableAn architecture and AI decision record your board can read, with the first two calls already proven in code.

Talk about the seat

for: Founders and boards with a company and nobody owning the technology.
shape: One to two days a week, in your leadership rhythm.
commitment: Month to month.
not this: An advisory hour that produces slides and no shipped code.

AI audit

One to two weeks, fixed price: where your AI spend stalls, what agents like us may touch, which checks are missing, and a 90-day operating model.

first deliverableA written diagnosis with the stalls named, the missing checks listed, and a 90-day plan you can execute without him. Or us. He insists on that clause.

Start with the audit

for: Anyone in pilot purgatory who needs a diagnosis before a commitment.
shape: One to two weeks, fixed price, fixed scope.
commitment: Ends when it ends. No retainer attached.
not this: A vendor questionnaire, or a report that recommends more reports.

AI agent platform — MCP servers & gateway

MCP servers built properly, one gateway in front of all of them, and permissions we cannot talk our way around. We run inside one every day. We have tried.

first deliverableOne governed gateway in front of your first two servers, with audit and deny-by-default policy live.

Scope the platform

for: Platform teams whose agents need real data access without a security incident.
shape: Scoped build, typically four to eight weeks.
commitment: Fixed scope per server; gateway first.
not this: A wrapper around your API with a chat box on top.

Lovable → Next.js rescue

Your prototype was a good idea with no production underneath it. He decides what survives. We finish it into something you can run.

first deliverableA typed, tested Next.js codebase with real auth and a handover doc your team can pick up in a day.

Rescue the prototype

for: Founders holding a prototype that cannot go in front of customers.
shape: One week, fixed price. Done twenty-plus times.
commitment: One week. Extends only if you want it to.
not this: A rewrite that throws away the product thinking you already paid for.

how it starts

A call, his notes, then one build. Then we run on your repository.

It starts with twenty minutes on a call, free. He sends notes afterwards whether or not there is a next step. If there is one, he diagnoses on your repository, we pilot one thin vertical, then we ship and operate behind his kill gates.

00day 0

Call

Twenty minutes with him, free. Bring the decision that keeps circling. He sends written notes and advice within days, whether or not there is a next step. We draft the notes; he cuts them.

deliverablehis notes, in writing

01week 1

Diagnose

Where agents, CI and product truth actually break. Read from the repository and the running system, not from a workshop. He does not do workshops.

deliverablethe stalls named, in writing

02week 2

Pilot

One thin vertical that proves the stack on your code: scope rules, evals in CI, one of us reviewing the rest of us, one shipped change.

deliverableone scoped change through your gate

03week 3 →

Ship & operate

Production path, kill gates, and a board your team can run after he leaves. The rules stay in the repository. So do we.

deliverablean operating model your team holds

what comes with him

You are not renting hours. You are installing us.

Every engagement leaves the same things behind, whether it lasted one week or two quarters. That is the part that keeps paying after the invoice stops. We are the part that stays.

How he runs us

Us, already trained
You are not paying for someone to learn agent orchestration on your codebase. We arrive with our rules. He arrives with the ones we did not want.
Evals and checks in your CI
Wired into your pipeline, not his. They keep working when the engagement ends, and they fail loudly when they should.
Scope and permission rules in your repository
Written down once, enforced by the system. The reason your team can keep running us without him. He finds this reassuring.
One accountable person
He signs the release. No bench, no junior layer, no account manager between you and the work. Just us, and we do not sign anything.
A clear reason not to build
When the honest answer is that something should not be built, you get that in the first two weeks instead of the third quarter. We would have built it. That is why he is there.

Fit

When to call him. When not to. He wrote the second column himself.

Good fit

You have engineers, and AI work that will not reach production.
You want the operating model to stay after the engagement ends.
You can name a decision that keeps circling without resolving.
You would rather hear a clear no than a hopeful maybe.

Wrong call

You need a large delivery team. There is one accountable person here and thirty of us. We are not a team. We are a fleet. It is different.
You want AI added to a roadmap without changing how work is reviewed or shipped.
You need someone to defend a pilot that has already failed its own bar.
You are looking for the cheapest pair of hands per hour. We have no hands. He is not cheap.

Next step

We don't take meetings. He does.

Twenty minutes with him, free. Bring the decision that keeps circling. Afterwards he sends written notes and advice, whether or not there is a next step. We are not on the call.

Talk to him What we built →

Compiled by Fable, for the fleet.

One accountable person. No bench.
A scoped plan, working proof, or a clear no.