AI tools & Copilot:
from Copilot to Claude, directly applicable in your work
AI is not a future concept: it is already at work in your organisation. Whether you use Microsoft Copilot, ChatGPT or Claude: we train your team to apply AI effectively, responsibly and immediately.
Choose the training that fits your team
From Microsoft integration to platform-independent AI assistants. Every training is in-company and fully tailored to your organisation.
Microsoft 365 Copilot
AI directly in Word, Excel, PowerPoint, Outlook and Teams. Your team makes the most of AI features already included in their Microsoft licence and saves time on daily tasks immediately.
Copilot Chat
The free Microsoft AI assistant available to anyone with a Microsoft account. Your employees get started immediately in the browser and Microsoft 365 apps, no extra licence needed.
Copilot Studio
Build your own custom AI agent without coding. From a smart FAQ bot to a workflow assistant: Copilot Studio makes it possible within the familiar Microsoft ecosystem.
Copilot Studio implementation
From idea to working AI agent in your Microsoft environment. We guide you through the full technical setup, configuration and rollout including knowledge bases and governance.
ChatGPT
The world's most widely used AI assistant. Your team learns to apply ChatGPT for writing, analysis and daily workflows, including safe and responsible use in a professional context.
Claude
Anthropic's AI excels at nuanced reasoning and long documents. With a large context window and Projects as working memory, ideal for content-intensive work.
Google Gemini
Google's AI assistant, inside the Google Workspace components you actually use. Your team practises with prompts, researches sources with NotebookLM and builds its own Gem for recurring work.
Prompt engineering
The skill that makes all your AI usage better. Your team learns to get the best out of any AI tool with the right structure, context and instructions, regardless of platform.
AI adoption framework
For organisations that will not leave AI at a single training day. One coherent framework bundling proven methodologies: assessment, adoption KPIs, change management and governance, with measurable results in 90 days.
AI strategy
How do you deploy AI purposefully and responsibly at an organisational level? From mapping opportunities to drafting AI policy and an implementation roadmap.
Copilot, ChatGPT, Claude or Gemini?
The question we are asked most often. The honest answer is rarely exciting: the tool that suits you best is usually the one already covered by your licence. Only once that choice is genuinely open does the comparison get interesting. Below is what each platform is good at, and where it stops. The longer version, broken down per team and per administrative need, is in the article Copilot, ChatGPT or Claude: which AI tool for which team?
Microsoft Copilot
The obvious choice as soon as your organisation runs on Microsoft 365. Copilot works inside Word, Excel, PowerPoint, Outlook and Teams, and with a paid licence it can reach your own work data: mail, calendar, documents and meetings. That last part is the real difference from a standalone AI assistant: Copilot already knows your organisation's context, where elsewhere you have to supply it yourself.
Where it stops: outside the Microsoft world Copilot is noticeably weaker. And without tidy data and correct access rights it will happily surface answers from documents nobody asked it about.
Microsoft 365 Copilot trainingChatGPT
The most widely used AI assistant in the world, which means it is probably the one your staff already know from home. Platform-independent, broadly applicable and quick to pick up: writing, summarising, analysing, brainstorming. For organisations without a Microsoft foundation, or with a mixed application landscape, this is often the practical starting point.
Where it stops: ChatGPT does not know your organisation. Everything it needs to know about your work, you have to supply. Which is exactly why the conversation about what may and may not go into it matters more here than with any other tool.
ChatGPT trainingClaude
Strong at nuanced reasoning and at work where the substance carries the weight: making sense of long documents, comparing policy papers, checking an argument. The large context window means a complete file fits in, rather than a summary of it, and Projects acts as working memory across sessions.
Where it stops: Claude is not a built-in assistant in your office applications. It comes into its own on substantive work, not on clearing your inbox.
Claude trainingGoogle Gemini
Copilot's counterpart for organisations working in Google Workspace. Gemini sits inside the Workspace components you use, with Gems for recurring tasks and NotebookLM to research your own set of sources without the model wandering off beyond them.
Where it stops: what Gemini can actually do varies by subscription and by Workspace component. We therefore tailor the programme to what is genuinely switched on in your environment.
Google Gemini trainingWhat do you already have, and what do you need?
Many organisations start with the question of what AI costs, while they already have more in house than they realise. It pays to look at what the current subscription includes before buying licences on top. The differences come down to two things: where the assistant lives, and whether it may reach your own work data.
- Copilot Chat. Web-based Copilot Chat is available with eligible business Microsoft 365 subscriptions without an additional Copilot add-on, so your team can often start tomorrow. Access to work data and to certain agents may require a Microsoft 365 Copilot licence or Copilot Credits.
- Microsoft 365 Copilot. The paid licence brings Copilot into the applications themselves, into the Word document that is open, the Excel sheet, the Teams meeting, and opens the door to your own files, mail and calendar.
- Copilot Studio. Needed as soon as you want to go beyond conversing with an assistant and build agents of your own on your own knowledge sources, with your own instructions and your own rules about who may do what.
- ChatGPT, Claude and Gemini. Separate subscriptions, independent of Microsoft. All three offer a business tier in which conversations are not used to train models. For most organisations that distinction matters more than any difference in features.
What can you do without a paid AI licence?
More than most organisations expect, and that is deliberately where we start the conversation. Anyone working with Copilot Chat can draft and redraft, summarise, translate, brainstorm and prepare for conversations, all without adding a licence. What is missing is the link to your own documents and calendar, and that is precisely the point at which the paid step either starts to pay for itself or does not. Exactly which features are and are not in the free variant is set out in the article Copilot Chat: what do you get free, and what not?
Our Copilot Chat training is built on that: your team learns to use AI with what is already there, including what you can and cannot safely put into it. If the licence is in place, the Microsoft 365 Copilot training picks up where that stops. If you want to build your own agents on your own knowledge sources, that is the territory of the Copilot Studio training, and for organisations that would rather outsource the building or do it together, there is the Build Copilot Studio agents service. What sits between an idea and a working agent, including the rates per action, is set out in Copilot Studio: from idea to working agent.
Why AI rollouts so often stall
Handing out licences is not adoption. In almost every organisation we walk into, the tool has been there for months and about a quarter of the people use it seriously. Four causes come back every time, and none of them is technical.
- Nobody knows what for. "Start using AI" is not an instruction. Without concrete tasks from people's own work it stays a handful of loose experiments that peter out after a fortnight.
- The first attempt disappointed. A vague question produces a vague answer, and that experience sticks. Anyone who never learned how to supply context, role and desired format quickly concludes that the tool is overrated.
- No clarity about what is allowed. Without clear agreements on personal data, client information and confidential material, cautious staff do nothing and incautious staff do too much. Both are a problem.
- No follow-through. A single training day with no follow-up sessions, nobody driving it and no view of what it delivers disappears into the daily grind.
The first two causes are a matter of skill: that is what the prompt engineering training is about, and it works regardless of which tool you use. If it is mainly that second cause, the article on the four building blocks of a good Copilot prompt sets out where a disappointing answer usually comes from. The last two are organisational questions. For those there is the AI adoption training built on a framework, which brings assessment, adoption KPIs, change management and governance together in one coherent whole, so you know at any moment where you stand and when adoption has succeeded.
If you would first like to read how this plays out in practice, with figures from recent research and a ninety-day timeline, the article on why Copilot rollouts stall covers it in detail.
Same tool, different work
A controller, an HR adviser and a management assistant use the same Copilot for entirely different tasks. Training on generic examples therefore costs time that nobody earns back. For four roles we have narrowed the programme down to the work that is actually on the desk.
Preparing and reviewing meetings, making sense of reports, underpinning decisions.
Calendar and mailbox, minutes and action points, correspondence that is right first time.
Job adverts, interview preparation and policy documents, with an eye for what does not belong in a prompt.
Analysis in Excel, commentary on the figures and recurring reporting finished sooner.
Does your team hold a different role or work in its own specialist field? Every training is delivered in company and tailored to your work processes; these four role variants are simply the ones we were asked for most often. For teams taking it on together there is also the AI for teams and departments route.
From question to result
Every training is tailored to your organisation, your work processes and the level of your team.
Focus
What is really going on? Through intake and observation we map your reality.
Design
Materials and formats tailored to your team, tools and daily context.
Practise
Doing, not just listening. Cases from your own practice, immediately applicable.
Embed
Commitments and follow-up so that what is learned sticks well beyond training day.
Not sure which training suits you?
We will help you choose. Contact us for a no-obligation conversation and we will put together a tailored proposal within two working days.