Live Claude enablement training
Your team already pays for AI. Most people are using it like an expensive search box. This training turns them into people who ship with it. Live sessions in plain English, built for the people who don't write code. Real work goes live every week, and when someone gets stuck, I unblock them on the spot. By the end, nobody is asking whether they can build something useful. They already have. The shape is simple: Level 1 makes users. Level 2 makes operators. Level 3 makes enablers. Here for yourself rather than for a team? There is a private executive track: the whole map, compressed and customized, one leader at a time.
The shape of it
Each level has its own job, and teams move up on real usage, not the calendar.
From first login to shipped work. By graduation, everyone has a working setup, an AI that knows them, and something real published that they built themselves.
People who drive the tool deliberately: they brief it like a colleague, control what it can touch, manage cost, automate their own work, and verify output with a method.
Champions inside your company who onboard new joiners, run the team's tool library, and keep adoption alive without an outside trainer in the room. That is the exit plan, and it is on purpose.
There is also a fourth door, described at the end of this page: a private executive track, the whole map compressed and fully customized, one leader at a time.
How it's run
A few principles do most of the work. They're why teams come out of this using Claude on their own, instead of nodding along in a workshop and going back to old habits the next morning.
Everyone hits small problems a video can't solve. I clear them the moment they happen, so nobody stalls. Troubleshooting is done out loud, as a teachable skill, because your team will hit the same walls again when I'm not there.
Every exercise runs on each person's actual job: their real documents, their real tools, their real backlog. Generic exercises get polite nods. Real work gets shipped, and shipped work gets used on Monday.
Each session ends with something built and shared, with a clear bar: good enough to post in your company channel. Nothing cements a skill like watching your own work go live and getting feedback on it.
Before we start
I send a checklist ahead of time and support it async, so nobody loses the first hour to installs and nothing installs live, ever. The last step is a verification prompt where Claude itself checks each person's setup and reports pass or fail.
Level 1 · Makes users
Four sessions plus the setup week, for teams new to Claude. The goal is simple. By the end, nobody is asking whether they can make something useful. They already have.
How Claude actually works, in plain language, and the moment that changes how people see it: everyone opens their AI's memory file live and finds it empty, then fixes that by talking to it for ten minutes and watching it write its own memory. Then the first build, from each person's real work.
Connecting Claude to the tools your team already uses, access keys handled with the one rule that actually matters, and the week's centerpiece: publishing real, on-brand work to the web at a link on your own domain. Branding comes last, content first, and a review pass strips the AI tells before anything ships.
The method that changes how people work. Run a task with Claude step by step once, verify each step, then turn the whole thing into a named skill and never give those instructions again. Everyone builds one from their actual week, and learns how to judge other people's skills before installing them.
The pieces become a morning routine: one command that pulls the day together, with the golden rule built in that nothing gets sent to anyone without human approval. Plus honest token literacy, the troubleshooting habits to keep going alone, and graduation demos into your company channel.
Level 2 · Makes operators
Six sessions for Level 1 graduates, with an entry bar: a working memory file, published work, and a skill in real use. This is where people stop making documents and start running systems.
Beyond "the file exists" into how to design it: a global file for who you are, project files that scope to their work, one purpose per file. An AI that genuinely knows your business is built here, deliberately.
Delegation as a written skill: brief the AI the way you'd brief a junior hire, then make it show its plan before it's allowed to act. Permission rules set deliberately instead of accumulated by fatigue-clicking, and the undo net demonstrated live, so caution stops costing capability.
Which model for which work, what a context window really is, and a live cost lab: the same task run wastefully and then well, with the difference measured in front of the room. Operators read their own usage instead of guessing.
A saved command, then a skill, then a scheduled run that drafts your Monday report before you wake up, then parallel workers for big jobs. Every rung with the same rule: drafts, approval gates, never a send button on autopilot.
The operator's daily bread: spreadsheet work in plain English, pointing Claude at a folder of similar documents and getting the comparison table, client-ready docs and decks, and browser automation for the tools that have no connector.
Client-data rules your team can actually recite, verifying output with a method instead of a feeling, and a capstone: one end-to-end workflow per person, combining data, processing, branded output, and an approval gate, demoed to the room.
Two electives slot in where a team needs them: GitHub for non-developers, the company filing cabinet for everything your team builds, operated entirely through Claude, and the meeting pipeline, which turns every meeting into extracted decisions, routed follow-ups, and memory instead of evaporation.
Level 3 · Makes enablers
Sold to the organization, not to individuals. One or two people per team, nominated by their manager, with protected time. Six sessions across twelve weeks, because the gaps are the curriculum: every session ends with a field assignment done with their real team, and by the midpoint the champions are running the room.
The facilitation craft behind these sessions, handed over: the rituals, the diagnostics, the two-minute rescue, the claim discipline. Then they run real mini-sessions with their own team, recorded, and we review the tape together the same way I review mine.
The skills library, curated and versioned. The onboarding runbook, so new joiners get a path instead of folklore. The working agreements: when AI-drafted work gets disclosed, how it gets reviewed, and who owns the error. Written by the team, facilitated by the champion.
Reading usage as a coaching signal, never a policing one. Handling the skeptic and the quiet non-adopter. Knowing what they handle, what goes to IT, and what comes to me. Graduation is a team playbook and a measured adoption result, not a certificate.
Champions keep a kit that stays current after the program: my delivery standards, session runbooks, a fact-checked claims register, and a monthly champions circle where it all gets refreshed. For the leadership layer there is also a 90-minute executive briefing: the decisions only leaders can make, from budgets and usage visibility to what should never be automated. It usually comes first.
The executive track · Private
For C-suite, founders, and business owners who want the capability themselves, without sitting in a team cohort. This is a private crash course through everything above, compressed and rebuilt around one person's actual role. No generic exercises exist on this track at all: every session runs on your real inbox, your real data, your real reporting, your real tools.
By the end you are not supervising an AI initiative. You are running your own AI operation, and you understand every piece of it because you built it with your own hands.
What's included, what they leave with
Live, department-specific sessions, run in rounds. Weekly office hours for follow-up and live problem-solving. Materials, exercises, and reference assets your team keeps. Progress tracked per team, with pacing set by how fast adoption actually sticks, and a written retro after every session so the program improves while it runs.
Real, on-brand work already shipped. An AI that knows each person and their job. Skills and automations they built themselves, with approval gates they understand. And past Level 3, something rarer: people inside your company who can keep all of it alive, which is the whole point.
$250 per hour, per session. Delivered department by department. Tell me which team goes first and we'll build the schedule around them.
