Skip to content
ELLmALet’s talk

Claude or Codex training · Swedish mid-market

Your people know
the business.
Now master
Claude or Codex.

Hands-on training for every department. Become an expert user of Claude or Codex and discover what AI can make possible for your business. I recommend starting with the leadership team.

Talk to Max Explore the training
4 sessions
Two hours each
5–9 people
A small, working group
Your work
No coding background needed
Max Lagrell, founder of Ellma and your trainer
Max LagrellYour trainer. From the first session to the last.

01 / What the research shows

High AI performance.
Visible leadership commitment.

AI high performers are twice as likely as others to report senior leaders’ commitment to AI initiatives. McKinsey includes role modelling as part of that commitment.

McKinsey, The state of AI in 2026: On the road to ROI
August 2026 · Report p. 19, Exhibit 11

02 / Start with the leadership team

See the potential.
Build the judgment.
Lead by example.

To make good decisions about AI, leaders need experience working with it. Using Claude or Codex deeply helps you understand its strengths, its limits and how it could change your business.

I recommend starting with the leadership team. Learn to delegate, collaborate and challenge the output. Build the expertise to recognise new possibilities and decide where AI belongs in the business.

Then lead by example. Show what good use looks like and help people across the organisation develop the same capability.

Built for

Established Swedish companies with around 50–250 employees. Every department is in scope. Leadership is the recommended starting point.

03 / In their own words

Real people. Their own workflows.

The work changes.
They can tell you how.

Christopher Sidgwick
Leadership · AI strategy
“Alight brought Max on board in 2026 to tailormake trainings for senior management in Claude Code, as part of our AI strategy. He's pedagogical, inspirational, and very knowledgeable - we can highly recommend him for training to excel in AI”
Christopher SidgwickHead of AI, Alight
Erik Isberg
Finance · Five group companies
“Eleven of us in finance, across five companies in the group, went through the programme. We now work in Claude against our own ERP, with workflows we built ourselves. Max has a rare combination: he is genuinely technical, he is good with people, and he pushes AI to its limits.”
Erik IsbergTapwell Group
Alexander Vase
Technical support · Alva Labs
“I connected Claude Code to Intercom, Google Cloud, BigQuery and our codebase. Now I run one Claude instance per support ticket, several in parallel. It pulls the conversation, the logs and the events, reads the code and drafts the reply. I don't read logs anymore. It saves me at least 10 hours a week.”
Alexander VaseTech Support Specialist, Alva Labs
Thobias Cogrell
Finance · Month-end close
“Before the training, I had never used Claude Code. Now the entire month-end close is automated. What used to take a week is now done at the push of a button, and I save around 30 hours a month.”
Thobias CogrellFinance Manager, Alva Labs

04 / The training

Four sessions.
A deeper understanding.
A new way to work.

Roughly half teaching, half hands-on work. Understand the technology, learn to direct an AI colleague and build skills around your own work.

Four two-hour sessions, with 5–9 people. We adapt the examples, workspace and tools to your group and whether you use Claude or Codex.

Tell me about your team
01

Foundation

See what expert use looks like. Start with working faster, then explore what AI makes possible. Shift from asking a chatbot questions to managing AI agents.

  • Chat, Cowork and Code, and where Codex fits into your work.
  • How your chosen tool works. When to delegate and when to collaborate.
  • A walkthrough of your personal AI workspace.

Hands-on workDocument your first workflow.

02

Training your AI colleague

Give your AI colleague the context and instructions it needs to work well with you. Turn your way of working into a skill you can use again.

  • The context window and what your AI colleague can see.
  • Workspace instructions, such as CLAUDE.md, and reusable skills.
  • MCP connectors: the connections you have and how to add more, with approved access.

Hands-on workBuild your first skill.

03

Understanding your AI colleague

Understand how the model produces an answer, where its knowledge ends and what happens when a long conversation is compacted. Use that understanding to direct and review its work.

  • Next-token prediction and the model’s knowledge cutoff.
  • Compaction and keeping the right context in a long conversation.
  • Loops: use the output of one skill as the input to the next.

Hands-on workBuild more skills and connect them into a workflow.

04

Working while you sleep

Explore how to give your AI colleague work that continues without constant supervision. Set clear instructions, boundaries and review points.

  • Subagents: divide work between several agents.
  • Routines: set up work to run again.
  • Loops: connect several skills into a longer sequence.

Hands-on workBuild more skills and put them to work together.

05 / Across the organisation

Training shaped around the people in the room

Every department.
Its own possibilities.

Sales, marketing, HR, customer support, product, finance, operations and beyond. The starting point is your expertise and what you want to do with it.

Understand

See what the
technology can do.

Explore Claude or Codex on work you know well. Learn how to give it context, where it is useful and where your own expertise matters most.

Explore

Go beyond the
way you work today.

Investigate a question, test an idea, build a tool or rethink a process. Use hands-on experience to see possibilities you might otherwise miss.

Put it into practice

Build capability
within the team.

Develop useful skills, review the results and share what works with colleagues. Keep improving how you work with your AI colleague.

Your context. Your systems, where access is approved. Your judgment on the result.

06 / Who you work with

I’m Max.
I build this way. I teach this way.

I ran three internal Claude Code cohorts at Alva Labs before taking Ellma full time. Today I help people across departments become expert users of Claude or Codex.

I work with people who know their jobs inside out. My role is to help them understand the technology deeply, explore what it can do for their business and build the confidence to work with it themselves.

Max Lagrell / Founder, Ellma

Let’s start with your team

What could AI
make possible for your business?

Tell me who’s in the group, how you use Claude or Codex today
and what you’d like to explore.

max@ellma.io