Why We Built This
The AI tools available in February 2026 are fundamentally different from what existed even six months ago. Claude Cowork, released in January, isn't a chatbot you ask questions—it's a worker you delegate tasks to. It produces actual deliverables: spreadsheets with working formulas, organised file structures, polished documents.
Most AI training hasn't caught up to this shift. Courses still focus on prompt engineering techniques designed for conversational AI. But when you're delegating outcomes to an agent that can work autonomously for hours, the skills you need are different.
We've designed a full-day programme that teaches what actually matters now: understanding how these systems work, providing effective context, delegating outcomes rather than describing processes, and working in files rather than chat.
The Demonstration-Led Approach
Our programme is demonstration-led, not lecture-based. That's a deliberate choice.
The problem with slides about AI is that they don't change behaviour. People leave knowing more about AI but not doing anything differently. Research consistently shows that hands-on practice with real tasks produces dramatically better outcomes than passive learning.
So we start with a demonstration—before any theory. A chaotic folder of files becomes an organised summary report in minutes. This isn't a promise of what AI might do; it's showing what it actually does, right now, with tools you can use today.
Throughout the day, every concept is illustrated with live demonstrations. Participants observe complete workflows, see real tasks transformed, and understand principles through watching them in action. Discussion follows demonstration, not the other way around.
Five Key Themes
The programme is structured around five themes that recur throughout the day:
Outcomes over process. The most common mistake people make with AI is describing how to do something rather than what they want done. "Go through each file, read the contents, identify the key themes, group them by category, and create a summary" is less effective than "organise these files into a categorised summary." Defining "done" clearly is the most valuable skill in working with AI.
Context is everything. The quality of AI output is bounded by the quality of context you provide. Rich input produces rich output. We spend significant time on how context windows work, what goes into them, and how to structure information for AI consumption.
AI-first, human-verified. Let AI produce the first draft. Your job is to review, refine, and verify. This is a mindset shift for many people, but it's the pattern that produces the best results with the least effort.
Preparation is the new execution. When execution is cheap—when AI can do the mechanical work—the value shifts to preparation. Thinking clearly about what you want becomes more important than the ability to do the work yourself. This is uncomfortable for people who've built careers on execution skills, but it's the reality of the current moment.
Files, not chat. The unit of work is no longer a chat message—it's a file. Input files, output files, deliverables you can use directly. If you're still copy-pasting from chat windows, you're using last year's workflow.
The Seven Sessions
The day breaks into seven substantive sessions:
Opening: The February 2026 Moment frames why this training, why now. We demonstrate what's possible before any theory, then set expectations for what participants will be able to do by end of day.
How LLMs Actually Work demystifies the technology without getting deeply technical. Understanding that LLMs are prediction engines—and why that leads to both their capabilities and their failure modes—helps people work with them more effectively.
Context: The Most Important Concept covers what context windows are, what goes into them, and why this matters more than clever prompting tricks. We demonstrate the same task with minimal context versus rich context—the difference in output quality is dramatic.
The Art of Prompting moves beyond basic prompt structure to meta-prompting (getting AI to improve its own prompts) and the evolution from prompt engineering to context engineering to delegation.
Files: The Unit of Work covers the paradigm shift from chat to files. Input files for grounding, output files as deliverables. We demonstrate Skills in action—Claude creating Excel spreadsheets with working formulas, PowerPoint presentations, structured documents.
Delegation and the AI-First Philosophy addresses the uncomfortable truth that AI is already better than most people at most knowledge tasks. This session covers delegation as a skill, quality verification when you're not an expert, and when not to use AI at all.
Advanced Patterns covers how the most effective users work: Skills for reusable procedures, sub-agents for parallel work, persistent memory with Claude.md files, and connecting AI to other tools.
The day closes with a synthesis session and interactive discussion applying what's been learned to participants' actual work.
What We Don't Do
Some deliberate omissions:
No hands-on exercises during the session. This might seem counterintuitive, but research on training effectiveness shows that demonstration-led learning followed by real-world application outperforms in-session exercises that use artificial examples. Participants observe and discuss during training, then apply to their actual work afterward.
No comprehensive feature coverage. We don't try to cover every capability of every tool. We focus on high-impact patterns that transfer across tools and situations. Learning everything is less valuable than learning the right things deeply.
No AI history or theory lectures. We don't spend time on the history of machine learning or the technical details of transformer architectures. Everything is in service of practical application.
No false promises. AI has real limitations. We discuss them honestly. Training that oversells capabilities creates users who either over-trust or eventually dismiss the tools entirely.
Who This Is For
The programme is designed for business professionals who need to use AI effectively in their work. It assumes basic computer literacy but no technical background. It's particularly valuable for:
- Knowledge workers whose jobs involve document creation, research, analysis, and communication
- Managers who need to understand what AI can and can't do for their teams
- Anyone who's tried AI tools casually but hasn't developed systematic approaches
- Teams adopting AI tools who need a common foundation of understanding
It's not designed for developers or technical AI practitioners—they need different training focused on different capabilities.
The Timing
We're offering this training at an unusual moment. The major tools just launched. Formal training hasn't caught up. The paradigm shift from "chat" to "work" is fresh enough to be genuinely surprising.
This means participants who go through the programme now are genuinely ahead of the curve, not catching up to it. That timing advantage won't last indefinitely—eventually the training industry will update and these approaches will become standard. But for now, there's a real opportunity.
The full programme details, including session-by-session breakdown and pricing, are available on our training page. If you'd like to discuss whether this is right for your team, get in touch.