There's a lot of tool building, but adoption remains a challenge. As Julia Mixter aptly put it during the forum, if AI is not aimed at the bottlenecks, it simply gets organisations to the same problems faster.

More throughput doesn't automatically mean more value. And more AI doesn't equal more adoption, unless it's aimed in the right place and people can feel its value in their day-to-day.

Capacity and value are not the same. If AI gives people time back, what happens to that time? Does it create space for better judgement and better service? Or just pressure to do more? 

If time saved simply becomes more work, the human cost of adoption needs much more attention. Have you asked yourself, what kind of workplace do you want to create with AI? 

Some use cases are already starting to feel essential. An example from Thirteen Group around Awaab’s Law stood out. 

When teams are working through high volumes, tight timescales, and complex repair categorisation, AI starts to feel less like experimentation and more like infrastructure. Not because it replaces judgement, but because it supports speed, consistency, and defensible action. 

The real governance challenge arises when AI stops arriving neatly. AI is now appearing inside systems people already use. New buttons. Embedded assistants. In-built functionality. 

That means governance and change initiatives can no longer rely on AI arriving as one clear initiative. People need to be ready before it shows up unannounced. 

There is a real tension between innovation and regulatory clarity. Housing is used to clear obligations, accountability, and regulation. AI in the UK is moving in a less settled landscape than that, which creates understandable tension. 

Not because there is no appetite to move, but because responsible progress is harder when the wider picture is still forming. Across these points, it's clear where the conversation needs to be. 

It's not about treating AI as a digital tool in isolation, but as a people, practice, governance, and behaviour challenge. Adoption does not fail on capability alone. It fails when the human conditions around it are left behind.