How we work · Train

We train your team.

Training on your own documents and tasks, in your office, then weekly office hours until your team is writing and sharing its own skills.

Who it's for

Leadership has decided AI matters, a few people have tried Claude on their own, and nobody's sure what to do next. The training answers that with your actual work, not a generic course. Every engagement includes it, whether it starts with a build or with the team.

What a session looks like

  1. A call before. We confirm the agenda, who attends and the examples from each team, so the exercises on the day are your own work.
  2. One or two days in your office. A three-hour morning and a two-hour afternoon, with the whole team in the room. We start with the basics: asking clearly, checking what comes back, and what confidential data stays out. Then your own work, team by team, in Claude and Cowork. Then, for anyone who wants to go further, Claude Code: small tools for the work your team repeats. Remote or hybrid when that works better.
  3. The deck, yours to keep. Written for your office, with the prompts and skills built on the day.
  4. Office hours after. An hour a week by video for the questions that come up on real work, until people are writing and sharing their own skills instead of asking us.

What it produces

  • A team that uses Claude on its own work. At Icelandic Glacial, finance and production went from first use to writing and sharing their own skills in weekly sessions.
  • A short list of what to move to Claude first, and why, from what we saw in the room.
  • Setup: accounts, access and configuration inside your own systems. Included in every engagement.

Where it leads

That list usually has a problem on it worth solving properly. When it does, we build it with the same team, and the training continues around what gets built. Lone Star trained every role first, then got the app. At IMRIS, an engineer in the pilot group went on to build the extraction tools himself.

Taught so far

IMRIS

A surgical imaging manufacturer's engineer built two extraction tools with Claude Code inside a controlled Microsoft environment. Case study

Icelandic Glacial

A bottled-water company went from scattered Claude accounts to one company-wide workspace, with NetSuite, Vividly and Kuebix connected. Case study

Lone Star Attorney Service

A Houston process-serving firm retired a 12-step manual onboarding routine for an app built with Claude. Case study

Pricing by proposal A first engagement runs about three months. It starts with a call.