Your AI Automation Roadmap

You already know that some part of your week should not be done by hand any more, and you have probably tried to fix it at least once. Then you sit down to actually do it and you are looking at twenty tools, forty tutorials and a hundred people telling you to go build an agent, and none of it is about your business, so the whole thing quietly goes back on the list for another month.

This call is for that exact moment. We spend two hours on the workflow you actually run every week, and by the end I can tell you what to automate first, what can wait and what you should leave alone completely.

It is 120 minutes on Zoom, just the two of us, and I record the whole thing so you are not sitting there taking notes instead of thinking.

The first half is me asking questions while you walk me through how the work really gets done today. Screen shared, real files, real mess, and no need to tidy anything up first. I am looking for where the hours go, which steps repeat every single week, which steps genuinely need your judgement and which ones only feel like they do.

The second half is the ranking, and this is the part people tell me they came for. We go through the candidates together and I tell you honestly which ones are worth building, roughly what each one will cost you in effort, and in what order they should happen so that the first one earns you the time to build the second one.

Then within three working days the written roadmap lands in your inbox.

The roadmap is a working document that you could hand to somebody else and they would know what to do with it. It has six parts.

1. A map of your current process and where the hours are actually going

2. Three to five things worth automating, ranked by time saved against how hard they are to build

3. For each one, which layer it needs and which tool actually fits. A prompt, a saved command, a skill, an agent, an MCP connection or an n8n workflow, running on Claude Code, Codex, an open model on your own hardware, Kimi or something else entirely. This is the part almost nobody gives you and it is the difference between a plan and a wish

4. What to ignore, and why. Usually the most useful page in the whole document, because an audit that recommends everything has quietly recommended nothing

5. The first build, specced out well enough that you or somebody you hire can start on Monday morning

6. The recording of our call, so you can go back to the bit where you said something clever and then forgot it

There is almost nothing to install for this one, and that is deliberate, because the whole point is to look at your work rather than at a tool.

What I do need from you before we meet.

1. Be ready to share your screen and open the actual sheets, inboxes, docs or tools where the work lives

2. Have one specific workflow in mind, the one that eats the most time. If you have three that is completely fine, we will pick together

3. Two clear hours with nothing stacked right behind them. This session runs long more often than it runs short

4. If other people touch the process, know roughly who does what

If you turn up without the second one we can still work it out live, however it costs us about twenty minutes that would be much better spent on the ranking.

A few honest pointers, because the wrong purchase helps neither of us.

If you already know what you want to build and you mostly want it built with you, then the better buy is From Blank Folder to Running System, which is two sessions of actual building rather than planning.

If you would rather learn all of this properly over six weeks with other people in the room, that is the Claude Code Builder cohort.

And if the answer turns out to be that somebody should build this for you rather than with you, we can talk about that too. In that case the $497 comes off the engagement, so this call never ends up costing you twice.

I keep 3 to 4 of these open a month, because the written roadmap takes real hours after the call and I would rather do four of them properly than ten of them badly.

Every booking includes a free 30 minute call with me, and it happens before the paid two hours. You do not pay anything extra for it.

We use it to work out what you are actually running and which tools are already in play, because that changes what the roadmap should say. It might be Claude Code, it might be Codex, an open model running on your own hardware, Kimi, n8n, or something you already pay for every month and have barely opened. I am not loyal to any of them, I have been wrong about several, and the ranking is a lot more useful when the tool question stays open until we have looked at your actual work.

Once that call has happened I set up the roadmap session itself and send you the invite.

If you would rather talk before you spend anything at all, the same 30 minutes is listed separately on my profile as a free AI discovery call. Book that one instead and we can work out together whether any of this is right for you.