Advanced Claude training for people who already use it every day and are ready to hand over whole tasks, not just prompts.
10,000+ people trained across 500+ organisations
Most teams plateau with Claude in the chat window. They write good prompts, get good answers and stop there. The real step up is Cowork and Claude Code: Claude working through a folder of files, running checks in parallel, building tools and dashboards, and doing jobs on a schedule while you get on with something else.
This hands-on session is for those confident daily users. We build it around your team’s own documents, models and recurring work, and we spend as much time on checking Claude’s output as on producing it, because that is what separates a useful agent from a confident mistake.
What people said after the class
100%
said the class will help them get more value from Claude at work
94%
enjoyed the class
93%
would recommend it to a colleague
5.4 hrs
average time each person expects to save every week
Post-class surveys from 32 attendees across four Advanced Claude classes at two global private equity firms.
Overall - AMAZING CLASS.
I have built a skill for a regular piece of work i do
Every session is built from scratch around your work, but every session builds the same foundations.
Setting up the Code tab in the Claude desktop app, choosing how much Claude may do before it asks, and having it interview you to write a CLAUDE.md file it reads at the start of every session.
Setting up a monitor that searches your sources every week and reports only what changed, with a source for every point.
Turning a messy document pack into clean text Claude can work with, and asking for quotes before analysis so every conclusion is anchored in the source.
Structuring prompts with clear sections and examples to answer things like investor questionnaires in your house style, then improving the prompt rather than correcting the answer.
Running the same check three times in parallel with independent sub-agents, comparing what they found and knowing that agreement is not proof.
Having Claude interview you to write a test for good work, turning it into a review Skill, and using it on a deck built with Claude in PowerPoint from an Excel briefing pack.
Building a dashboard from a workbook where the data rules are written down and run as code, so errors are flagged rather than smoothed over.
Restructuring Skills so Claude only reads what it needs, testing that each one triggers on the right requests, and choosing the cheapest model that still passes your test.
The capstone
People finish by pointing what they have learned at their own work: a sourcing agent that checks its own evidence, a shared team brain or a scheduled chief-of-staff briefing that drafts but never sends.
We can usually map your highest-impact use cases in a single conversation.
Get in TouchThese came out of our advanced sessions. Each one hands a whole job to Claude and builds in the checks that make the result safe to rely on.
Compares consecutive monthly board packs to surface contradictions, restated figures and changed definitions with page references, run several times independently, then lists questions for management.
Searches public sources for companies that match the fund’s criteria and deal signals, cross-checks the CRM and uses a separate sub-agent to verify every source before a name reaches the shortlist.
Each weekday morning, triages email and calendar, drafts replies in your voice, briefs external meetings and flags overdue commitments, without sending anything.
Combines deal newsletters in the inbox with web search into a ranked weekly roundup of deals, leadership changes, profit warnings and regulation, each with a source link.
Drawn from our anonymised client use case library.
Three signals from Anthropic’s own research on how Cowork is being used. Each links to its source.
Advanced sessions are shaped around one team’s work, so every exercise uses documents and tasks that look like theirs.
01
A call to understand your team, the connectors and tools your IT team allows, and the work people most want to hand over.
02
We build realistic invented files, such as management accounts with planted errors, so people practise on work that looks like theirs without sharing anything confidential.
03
A hands-on session that moves from setup to advanced techniques, ending with stretch goals people choose for their own work.
04
A follow-up call to see which Skills, agents and routines made it into everyday use and what to build next.
Delivered in person at your offices, or remotely.
Case study: Access Holdings
Access Holdings, a private equity firm in Baltimore, trained 27 members of its deal team on Claude with us, then brought operators from across its portfolio together for a full-day AI value creation offsite.
Going further
Once people are using Cowork and Claude Code day to day, the next step is turning individual wins into shared ways of working.
People who already get good results in chat and want to hand over bigger, multi-step jobs.
Private equity, growth and fund teams working through board packs, models, questionnaires and sourcing.
Advisers, legal, compliance and tax teams who need speed without losing the audit trail.
The people others already ask for help, who will set the standard for how the wider firm uses Claude.
Other courses teams pair with this one.
Every Claude Cowork session is led by one of our instructors, with a dedicated tutor in the room so nobody gets left behind.
Founder + Chief AI Officer
AI Instructor
AI Instructor
AI Instructor
AI Instructor
AI Instructor
AI Instructor
AI Instructor
Mastering Claude gets a whole team confident with Claude in the chat window. This session assumes that is done. It moves people on to Cowork and Claude Code, sub-agents, Skills, scheduled tasks and writing their own tests, so they can hand over whole jobs and still trust the result.
Cowork is Claude’s agent for longer, multi-step work: reading and creating files, using your connected apps, running checks in parallel and doing tasks on a schedule. Anthropic is now folding it into the main Claude app, so it is less a separate mode and more how Claude gets real work done.
No coding is needed. People should be comfortable using Claude every day. We use the Claude desktop app throughout, and we send your IT team a short setup checklist before the session.
No. We build realistic invented documents and workbooks that look like your team’s work, often with deliberate errors to catch, so nothing sensitive is needed.
In person at your offices by default, with an instructor and a dedicated tutor in the room. We can run it remotely if that suits your team better.
Yes. We are based in Covent Garden, London, and run most sessions in person at clients’ offices in London and across the UK. We also work with teams worldwide, from Paris and Berlin to New York and Singapore, and can run sessions remotely when that suits better.
From The Distillation, our newsletter on putting AI to work.
GPT-6 Astra and Claude Opus 5.5 moved the frontier twice in three weeks. Plus AI training for the industries that make things, field notes from our US rollouts with Access Holdings portfolio companies, and funded graduate places on our next open enrolment sessions.
Josh Page ·
An Anthropic researcher quit to warn us about the risks. People keep asking me how worried we should be. The answer is quite, and the reason isn't last week's headlines. It's what happened at Hugging Face eight weeks ago.
Tom Hewitson ·
Access Holdings didn't wait for their portfolio companies to work it out. Other firms should take note.
Tom Hewitson ·
Tell us how your team already uses Claude and the work they most want to hand over, and we will design an advanced session around it.
Get in TouchSources and further reading