AI-NATIVE COMPANIES DON'T ADD LAYERS

Did your mother nag you about the virtues of wearing layers? Turns out they might not be as useful when it comes to making your enterprise AI native.

A new study concludes that firms built around AI, from the very start, are organised differently. Not only do they reach the same valuations when 25% smaller, they have more senior workforces, fewer managers and flatter hierarchies.

But - perhaps most insightfully for regulated enterprises - they haven’t had to add AI layers over existing workflows.  Their advantages come from building AI into what the business is making. At a fundamental level.

Which means digital intelligence gets to do its thing. And human judgement adds the most amount of value possible. Digital and human intelligence is integrated to a far greater degree.

Of course, this is simpler for a startup to achieve. But it is the end state that all of us, even those in regulated and complex industries, need to aim for. Because every extra layer of integration increases dependency and adds unnecessary friction to workflows.

It will still mean that a deviation lands in quality at 7.40am. The batch record will be open. A sensor trace will need checking against the run history.

But the AI will read the record, compare the run, surface the exception, draft the rationale, show its evidence and leave a trail that a qualified person can challenge.

And the deviation will be solved far more quickly than it would have been if layers of AI and layers of human decision making had to interact and rub up against each other along the way.