The question is not whether to use AI. It is how.
AI is displacing work faster than it is creating it, and around 2.3 million people in the UK are already at risk of being left behind. Kritmatta's approach starts from a different question: not whether to use AI, but how to build it so people can actually use it.
Most conversations about AI in business still start with the wrong question.
Should we be using AI? It gets asked in board meetings, in strategy documents and roughly nine times a day on LinkedIn. It is also already out of date. The technology is in the building whether or not anyone signed it off, being used by individual employees on individual tasks, usually without a policy anywhere near it.
The more useful question is the one underneath: how do we use it, and who gets to use it.
The 2.3 million people missing from the conversation
Around 7 per cent of the UK's eligible workforce is classed as digitally illiterate. On ONS figures, that is roughly 2.3 million people who could be contributing more to the economy than they currently are.
That number rarely appears in AI conversations, because most AI conversations quietly assume a baseline of digital confidence. They assume someone comfortable enough to try a new tool, articulate enough to prompt it well, and senior enough to be given access in the first place.
Strip those assumptions away and the picture changes. The people most exposed to AI displacement are frequently the least equipped to benefit from AI adoption. A tool that requires technical fluency to operate does not close that gap. It widens it.
Enabler is a design decision, not a slogan
Plenty of companies describe their AI as an enabler. Far fewer have made design decisions that would actually make it one.
If the goal is that nobody gets left behind, it has to show up in the product itself. That means no code, because requiring code immediately excludes the majority of the people who would benefit most. It means describing what you need in plain language rather than assembling a flowchart. It means working with the tools a team already uses, because migration is a barrier disguised as an upgrade.
It also means everything being visible. Every action logged, every decision traceable back to the message or document it came from. Trust is not a feature you add later. It is the thing that determines whether someone uses a tool twice.
None of that is idealism. It is the difference between a product that gets adopted across a team and one that gets used by the two people who were already comfortable with technology.
What the last technology shift actually looked like
There is a version of the AI conversation that skips straight to inevitability. The jobs are going, the argument runs, so the only question is how quickly.
History is more mixed than that. When farming mechanised, and when the railways moved from steam to electric, work was displaced and work was also created. The disruption was real and so was the recovery.
The honest position on AI is that only half of that has happened so far. The displacement is visible. The new roles have not arrived at the same pace, and pretending otherwise helps nobody.
That does not make the doomsday scenarios right either. The idea that the only remaining option is universal basic income assumes the creation side never comes. There is little in the record to support that, and a great deal of noise in the current debate that has more to do with the economy and policy than with the technology itself.
What it does mean is that the gap between displacement and creation is where people actually get hurt. Which is precisely where enablement matters most.
The bar we are setting
At Kritmatta the aim is for AI to be an enabler, and to do our bit to make sure nobody gets left behind on the AI transformation journey.
In practice that is a narrower commitment than it sounds. It means the person who has never written a line of code should be able to build something useful. It means the tool should reduce the amount of work someone does, not the number of people doing it. And it means the judgement calls stay with the human who has the context, because context is the one thing an AI system does not have.
The question was never whether to use AI. It was always how, and for whom.
Simple AI. Real results.