Most AI conversations start in the wrong place.
People obsess over prompts.
What’s the perfect prompt?
What’s the framework?
What’s the magic formula?
But we’re focusing on the wrong problem.
The organisations and individuals getting the most value from AI aren’t necessarily writing better prompts. They’re providing better context.
And that’s a fundamentally different challenge.
Because if prompts are instructions, context is understanding.
Without context, AI can generate content.
With context, it can become genuinely useful.
We’ve moved beyond prompt engineering
For the past two years, prompt engineering has dominated AI conversations.
Understandably so.
When a new technology arrives, our first instinct is to learn how to operate it.
But as AI tools become more capable, the differentiator is shifting.
During the discussion, Toby described how a recent client project completely changed how Freeformers delivered its work. Rather than receiving a polished presentation deck or traditional consultancy report, the client requested something unexpected: a markdown file designed primarily for AI systems to read and interpret.
At first glance, it sounds technical.
But the implications are much bigger.
We’re beginning to move from prompt engineering towards context engineering.
Not focusing solely on what we ask AI to do.
Focusing on what AI understands before we ask.
The real problem isn’t intelligence. It’s understanding
One of the most common criticisms of AI is that it sometimes produces generic outputs.
The response sounds plausible.
It sounds professional.
But it doesn’t sound like you.
That’s because AI is fundamentally a pattern-recognition system.
It doesn’t inherently know your experiences, beliefs, communication style, career history, decision-making process, or perspective unless you provide that information.
As Toby pointed out, the goal isn’t simply creating content that sounds human.
The goal is creating content that sounds like you.
That’s a very different objective.
And it requires a very different approach.
Your most valuable AI asset already exists
Most people are sitting on years of valuable context without realising it.
Your CV.
Your LinkedIn profile.
Your presentations.
Your proposals.
Your blogs.
Your podcasts.
Your emails.
Your career history.
The challenge isn’t creating more information.
It’s organising what already exists.
Toby shared how he built what he jokingly refers to as “Toby GPT” by combining years of written content, presentations, conversations, podcasts, proposals and professional experiences into a usable knowledge base. The result wasn’t simply an AI that could generate content. It became a system that understood how he thinks, writes and communicates.
That’s where context becomes powerful.
Because context doesn’t just improve output quality.
It preserves nuance.
Start with yourself before you start with your organisation
One of the most practical insights from the episode was also the simplest.
Don’t start by building an enterprise AI strategy.
Start with yourself.
Toby’s recommendation was surprisingly straightforward: create a structured document about your career and professional experience. Whether that’s a markdown file, a Google Doc, a Word document or a collection of notes matters less than people think. The important thing is creating a source of truth that helps AI understand who you are.
Emilie was keen to stress that this doesn’t require technical expertise.
You don’t need to understand markdown.
You don’t need to be a developer.
You don’t need a sophisticated knowledge architecture.
You simply need usable context.
The technology is becoming increasingly accessible.
The bottleneck is no longer the tool.
It’s the information we provide it.
The future belongs to builders, not users
Another theme running through the conversation was the distinction between using AI and building with AI.
Many people currently use AI to rewrite emails.
Summarise documents.
Generate meeting notes.
Speed up existing tasks.
And that’s useful.
But it’s only scratching the surface.
The bigger opportunity is redesigning how work gets done.
Not asking:
“How can AI help me do this faster?”
But:
“How would I design this differently if AI existed from the beginning?”
That’s a much more transformative question.
And it’s the mindset separating experimentation from innovation.
The human advantage isn’t disappearing
For all the discussion around AI capability, one of the most interesting moments came when Emilie asked Toby what he would never outsource to AI.
His answer wasn’t strategy.
It wasn’t writing.
It wasn’t problem-solving.
It was listening.
Specifically, listening to people.
The ability to sit in a conversation, observe behaviour, understand emotion, recognise nuance, and make connections that aren’t explicitly spoken.
As practitioners of human-centred design, Toby and Emilie highlighted something many organisations risk forgetting in the rush towards automation:
Data tells you what happened.
Humans often understand why.
And in a world increasingly driven by AI, that distinction becomes more valuable, not less.
AI adoption isn’t really a technology challenge
Towards the end of the discussion, the conversation shifted to an article about AI adoption inside a large organisation.
The headline claimed the biggest blockers weren’t technical.
They were human.
That observation feels increasingly true.
Most organisations don’t have an AI problem.
They have a behaviour problem.
A learning problem.
A prioritisation problem.
People are busy.
Work already fills their calendars.
Finding time to rethink how work gets done can feel harder than simply carrying on as before. As Toby noted, stopping to redesign your workflow often feels like it gets in the way of doing your actual job.
Which creates a paradox.
The very activity that could make work easier is often the thing people feel they don’t have time to do.
Context is becoming a strategic asset
The organisations that thrive in the AI era won’t necessarily be the ones with access to the most powerful models.
Those tools are rapidly becoming available to everyone.
The differentiator will be context.
Who understands their people best?
Who structures knowledge most effectively?
Who captures expertise before it leaves the organisation?
Who creates systems that help technology understand how their business actually works?
Because AI without context is generic.
AI with context becomes capability.
And capability creates competitive advantage.
The question every leader should be asking
Not:
“Which AI tool should we buy next?”
But:
“What knowledge, experience and expertise do we already have that we’re failing to capture?”
Because the future of work won’t be won by organisations with the best prompts.
It will be won by organisations with the richest understanding of themselves.