The Maturity Ladder

Over the last year we've done what most organisations rolling out AI have done. We've released tools, encouraged experimentation and given people space to work out where they're useful.

That experimentation matters. Nobody starts by building organisational solutions. People start by solving problems for themselves. They summarise meetings, draft documents, analyse data and automate the irritating parts of their day. The value is immediate because it's personal.

For a while that's enough.

Then eventually the same question appears.

"This is useful, but where does it go?"

Not because personal productivity lacks value. Quite the opposite. Most of the interesting work seems to start there. The challenge is that personal productivity was never the destination. Organisations don't invest in these capabilities simply to create isolated individual gains. At some point those gains need to become something larger. The question becomes how you move people from personal productivity towards the kinds of use cases capable of creating organisational efficiency.

That's the problem we found ourselves trying to solve.

We had no shortage of ideas. People were finding use cases everywhere. What we lacked was a reliable way of describing which activity was moving somebody closer to organisational outcomes and which activity was simply making an individual more effective.

One of the reasons we aligned on this thinking relatively early was because we kept seeing a similar pattern emerge in the more successful ideas. Not enough examples to pretend we'd discovered a universal law, but enough examples to believe there was something useful underneath them.

Looking back, the journey appeared to follow a recognisable shape. It often started with prompting, became something reusable, evolved into a personal agent and then into a shared team capability. Beyond that, the centre of gravity shifted away from personal productivity and towards organisational outcomes through end-to-end agents and, eventually, applications.

The exact stages aren't particularly important. Different platforms will have different feature sets. What mattered to us was that beneath dozens of individual features there appeared to be a much smaller skeleton. A handful of capability shifts that consistently seemed to move people from personal productivity towards organisational efficiency.

Once we had a way of describing that skeleton, a number of things became easier.

Firstly, it gave us a way of understanding use cases. A prompt and an application aren't simply different solutions. They're different positions on the same pathway.

Secondly, it gave us a way of understanding capability. If people sit somewhere along the pathway, then capability can be assessed against the pathway rather than against an ever-expanding list of product features. Training also becomes easier to design because the focus shifts from teaching features to developing the capabilities needed to move to the next stage.

Most importantly, it provides a way of measuring progress.

The objective isn't to prove that organisational efficiencies already exist. In many cases they won't. The objective is to understand whether capability is moving in the direction most likely to produce them. If organisational efficiency sits further along the pathway than personal productivity, then movement along the pathway becomes a meaningful measure in its own right.

That same logic scales surprisingly well.

Assess individuals and you get an individual maturity read. Roll those assessments up and a team pattern begins to emerge. Roll them up again and you start to see organisational capability. Alongside that sits a second assessment, which is organisational readiness itself. Infrastructure, governance, security and platform capability determine how much of the pathway can be safely supported regardless of how capable people become.

Taken together, those two reads provide a practical way of running the programme. One describes where capability exists. The other describes what the organisation is currently able to support.

We didn't set out to build a maturity model.

We were trying to answer a much simpler question: how do organisations move from personal productivity to organisational efficiency?

In trying to answer that question we found ourselves defining a smaller skeleton sitting underneath a much larger set of AI features. That skeleton provides a way of describing use cases, assessing capability, structuring development and understanding readiness through the same lens.

Most importantly, it creates a measurable way of understanding whether progress is moving in the direction that matters.

The challenge isn't releasing more features.

The challenge is helping people move through the capabilities that matter, whilst ensuring the organisation is ready when they get there.