Claude Code Intensive Training
Turn employees into AI Experts
Turn up with your laptop. Leave with the toolset to build. This is an interactive Claude Code masterclass for non-technical teams.
Proof
What participants say
Audience
Who this is for
Marketing
Create agents to generate unlimited on-brand content. Add a connector to pair insights with deliverables. Build custom apps to delight stakeholders.
Design & creative
Don't wait for engineering. Create full-stack applications from scratch and orchestrate agents to implement features on existing apps. Use Claude Code as mission control for your entire design methodology.
Sales
Prepare account research, tailor proposals and outreach from your own templates, and tidy CRM exports. Standardise the repetitive parts of the pipeline into steps Claude Code runs on request.
Overview
Workshop overview
Claude Code is an agentic AI coding assistant created by Anthropic. It runs as a desktop app, inside an editor such as VS Code, or in the terminal - reading and editing files, and working through multi-step tasks on its own. This series teaches non-technical teams - marketing, design, sales, and finance - to use it directly, with no coding background assumed.
Teams learn to do the work they currently wait on engineering for, making the organisation more productive take on more without increasing headcount. Every exercise is built on the team's own files and data, so the skills apply to real work from the first day. Because each day builds on the last, spending on AI becomes reusable capability rather than a one-off result.
Curriculum
The three days
The days are cumulative and run in order. Each builds on the one before, so a team starts at Day 1 and continues as far as it needs. Every exercise uses your team's own tasks and files, so the series fits whatever your function does.
Fundamentals: drive Claude Code on real work
Setup on every surface — desktop app, VS Code, terminal — then straight into real work. Participants learn how Claude Code actually operates: it reads your files, makes changes, and asks permission before acting. The core skill of the day is prompting — giving it the right context, scoping requests so they succeed first time, reviewing every change in plain terms, and undoing anything. Each person finishes a task they currently hand off or wait on.
By the end of the day
Everyone finishes a real piece of their own work — and can direct, check, and undo Claude Code without help.
- Set up Claude Code and learn how it works: it reads your files, makes changes, and asks first
- Prompt with context — point it at files, paste screenshots, scope requests so they land first time
- Review every change in plain terms; use checkpoints and rewind so nothing is ever stuck
- Have it plan and explain its approach before it touches anything
- Finish a real task — document, report, or brief — before the day ends
Power features: make it remember, make it repeatable
Day 2 covers Claude Code's memory and automation features — so the team stops re-typing what the tool should already know. Tone, terms, and standards go into CLAUDE.md once and apply automatically for everyone. Repeated tasks become custom slash commands and skills anyone can run. Plan mode handles the bigger jobs, and the team's work moves into GitHub — every change reviewed, reversible, and deployed automatically on approval.
By the end of the day
The team's standards live in CLAUDE.md, its most-repeated tasks run as one-command workflows, and its work is version-controlled — reviewed, reversible, deployed on approval.
- Write your tone, terms, and standards once in CLAUDE.md — applied automatically, for everyone
- Save repeated tasks as custom slash commands and skills the whole team shares
- Use plan mode and extended thinking for jobs too big to wing
- Keep long work healthy across sessions — resume yesterday's thread, fix work in sections
- Ship through GitHub: Claude Code writes the commits and pull requests, you review, deploys run automatically
Advanced: agents, connections, and autonomy
The features that make Claude Code an agent platform rather than an assistant. The team connects it to the tools and data it already uses with MCP, defines subagents that handle specialist jobs in parallel, and uses hooks and headless mode to run work on schedules and triggers — overnight, on deadlines, when files land. The capstone uses all of it: each participant builds and deploys a working application from scratch.
By the end of the day
Agents run the team's routine work around the clock on its real tools, governed by rules it wrote — and every participant has built and shipped a working application.
- Connect Claude Code to the data, tools, and systems your team already works in with MCP
- Define subagents for specialist jobs — research, drafting, checking — and run them in parallel
- Automate with hooks, schedules, and triggers — work that runs overnight or when files land
- Build and deploy a complete working application from scratch
- Set permissions and sign-off rules IT can approve — what runs unattended, what doesn't
Workshop details
Each day is £5,000 for up to 25 participants. The series is cumulative: teams start at Day 1 and continue as far as they need.
"The best part of the workshop was realising that AI can actually take a task off your plate entirely. I left feeling confident that I finally have a practical way to use AI in my day-to-day work, not just theory or hype."
