The operator-led agent studio.

For early-stage teams that use agents every day and don't get much back.

Most teams are just bolting agents onto the work they already do. Take a messy workflow, add an agent on top, and now you have a messy workflow with an agent on top.

The teams that win do the opposite. They see that the old workflows are shaped for humans, and redesign them from scratch for agents, now that intelligence is cheap.

That's the hard part. Not building the agent, choosing the workflows worth rebuilding, and getting your team to trust them. That's what I do.

I'm Ryan. I was the first employee at a fintech startup, from seed to Series A. I'm an operator who learned to build, not an engineer who picked up business on the side.

The question now is not can we build it. It's should we.

So I made the first part free.

Run the Workflow Audit, inside your coding agent.

It interviews you about the workflows that eat your week: tracking what competitors ship, chasing sales call follow-ups, or repurposing content.

Then it hands you a report you keep:

  • where your old workflow breaks
  • which parts to automate
  • which need your judgement
  • what to build first
  • how to know you can trust it

It diagnoses before it suggests. If a workflow isn't ready to build, it tells you so, and tells you why.

Works in any coding agent. Paste this into a new session:


            

This page has an agent-readable version: the note and the install prompt, exactly as your coding agent receives them.

If it finds something worth building, I build it.

One workflow takes about a month, with human review and evals built in.

I stay in the loop after it ships, because the work isn't done until your team trusts it.

Builds start at £4,500 for one workflow built and adopted. You only pay once there's a workflow worth redesigning.

One honest note.

This is v0. No case study, logos, or chart that goes up.

What you get is me: I map your workflow, build it, and stay on the hook for whether your team actually uses it.

Questions

Nothing. You copy a prompt into your own coding agent and run it: no form, no call, no sign-up. You keep the report either way.

Then it tells you so, and tells you why. It diagnoses before it suggests, and it will say a workflow is not worth building yet rather than sell you one. That is the point: you trust the yes because it can say no.

Builds start at £4,500. That buys one workflow mapped, built, and deployed with your team, with human review and evals designed in, over about a month. You only pay once there is a workflow worth redesigning.

Any of them: Claude Code, Codex, Cursor, or a plain chat. Claude Code installs in one line. Everywhere else you point the agent at the skill files and it runs from those.

No. The audit interviews you about a workflow you already run by hand. I am an operator who learned to build, so the build matches how your team works rather than adding a stack you have to maintain.

Want a second opinion before you run it? Ask an LLM what In The Loop does, and whether it fits your team.


Pick the workflow that's been quietly eating your team's week, and run the audit on it. It costs nothing, and you keep the report either way.