Partner Article

Whitehall has reorganised itself around AI. Now someone has to teach people to use it.

By Paul Smissen, Director at We Are Zenith

What the Burnham reshuffle means for employers in the North East, and the bit that keeps getting skipped.

Without getting political, and I’ve really tried not to, I’ve been following Andy Burnham’s first few days in the job pretty closely. Particularly what he’s done with technology and AI, because that lands directly on the work we do every day.

This is my first proper go at writing something like this. It isn’t a policy analysis and I’m not pretending to be a Whitehall watcher. It’s a view from where I sit, which is in front of North East employers most weeks, talking about skills, people and increasingly about AI.

What actually happened

For anyone who missed it, here’s the short version.

The Department for Science, Innovation and Technology has gone. Set up by Rishi Sunak in 2023 to put science and tech at the centre of government, it lasted about three and a half years.

Its science and innovation work now sits inside an expanded business department, renamed the Department for Business, Innovation, Science and Trade, led by Jonathan Reynolds. Digital policy and online safety have gone back to culture, which picks up the D in its name again and becomes DDCMS under Lisa Nandy. AI strategy, public sector AI adoption and the AI Security Institute have all moved into the Cabinet Office.

And Kanishka Narayan, the AI minister, now has the right to attend cabinet, in a post that sits jointly across the Cabinet Office and the new business department. That’s a first for the UK, though not globally. The UAE appointed an AI minister back in 2017, and Canada put one in full cabinet last year.

The reasoning behind it is straightforward enough. You can’t run a coordinated national industrial strategy with science and technology sitting off to one side in a department of their own, slightly apart from the businesses meant to do something commercial with any of it. Put them in with trade and industry and you shorten the distance between a good idea and an actual job.

The reaction has been mixed, and fairly

The tech sector hasn’t taken it well, and the objections landed before the announcement even did.

techUK and the Startup Coalition wrote a joint open letter to Burnham the Saturday before he took office, calling the plan “the wrong change at the wrong time” and arguing that scrapping the department sends the wrong signal to an industry growing at around 10% a year. Their case was that DSIT gave the sector a single front door into government. Dom Hallas of the Startup Coalition put it more bluntly, warning that in a merged department British tech would end up competing with British steel for ministerial attention. Ian Hogarth, who chairs the AI Security Institute, reckons a reorganisation on this scale distracts a department for the best part of a year, and pointed out that good civil servants tend to leave when their time gets eaten by restructuring.

Paul Nurse at the Royal Society is making a different point. His worry isn’t pace, it’s status, and whether science ends up the junior partner in the new arrangement. Reynolds has spent the days since insisting it won’t.

Worth saying that the pressure hasn’t all run one way. Digital rights groups including Open Rights Group wrote to Burnham the same week arguing he should go further in taking on the big technology and AI firms, not less far. So “the sector is unhappy” is really “the trade bodies and the startup lobby are unhappy”, which isn’t quite the same thing.

The voice that caught my eye was Russ Shaw of Global Tech Advocates. He made the point that Burnham’s instinct to spread growth beyond London could genuinely strengthen regional tech clusters, provided it doesn’t come at the cost of a coherent national plan. He’s the only one I’ve seen making the regional argument at all, and for obvious reasons that’s the version of this I’m most interested in.

I’ll be honest with you. I don’t have strong feelings about which building a policy team sits in.

I’ve watched enough reorganisations to know the letterhead changes long before anything else does. What happens in a Newcastle manufacturer or a Teesside logistics firm next Tuesday isn’t decided by whether the acronym is DSIT or DBIST. It’s decided by whether anybody in that business has the confidence and the permission to work differently.

It’s the other half of the announcement that I think matters more.

A seat at the table is a signal, not a strategy

Putting AI in the cabinet room says government has stopped treating this as a niche departmental interest and started treating it as an economic question. That’s the right call, and I’d say the same whoever had made it.

Narayan has made his own case plainly enough. The best version of this, as he sees it, is a reindustrialised Britain, stronger national security, and public services that work better than they do now.

I want that too. Genuinely. But there’s a step in the middle that keeps getting skipped, and it’s the step we spend all our time on.

The barrier isn’t the technology

Nearly every business we work with could have Copilot or ChatGPT switched on by Friday afternoon if they fancied it. Plenty already have. Access is not the problem, and hasn’t been for a while.

The problem is what happens next, which is usually not much.

Here’s the pattern, and I’d guess a lot of you will recognise it. The licences get bought, often on the back of a good demo. A launch email goes round. Maybe there’s an hour’s training, delivered generically, covering features rather than the actual job anyone in the room does. A handful of enthusiasts use it properly. Everyone else tries it twice, doesn’t get a great answer, quietly goes back to how they did it before, and three months later it’s another tool gathering digital dust.

Nobody changed how the work actually gets done. Nobody gave the people doing it the time, the permission or the confidence to work it out for themselves. And crucially, nobody measured anything, so there’s no evidence to point at when the renewal comes round.

Ship the technology without the capability and you don’t get a productivity gain. You get a subscription.

What the businesses getting this right do differently

They aren’t doing anything exotic. In my experience it comes down to four things.

They start with a job, not a tool. Not “let’s roll out AI”, but “quoting takes us three days and it shouldn’t”. The tool gets chosen after the problem is named, not before.

They give people protected time. Capability doesn’t appear in the gaps between other work. The firms that get somewhere carve out actual hours, and their managers defend them.

They build it into someone’s role. Not a committee. One or two people whose job description genuinely includes making this work, with the authority to change a process rather than just recommend one.

They track something. Even roughly. Hours saved, error rates, turnaround times. Not for a dashboard, but so the next decision is made on evidence rather than enthusiasm.

None of that requires a technology department. All of it requires people who’ve been developed to do it.

What this means for us, and the challenges I can see

If government is serious about pulling science, innovation and business into one place and calling it an industrial strategy, then skills stop being a nice-to-have and become the delivery mechanism. That’s the space we sit in. Recruit, develop, retain, with AI capability increasingly running through all three.

So on the whole, I think this is good news. But there are two things I’d keep an eye on.

The first is reach. The businesses that stand to gain most here are the 40-person outfits with no IT department, no innovation budget and an MD already doing three jobs. They’re never in the room when this stuff gets designed. That’s not a dig, it’s just how it tends to go. And if a national AI push only lands with the firms who already had a technology function, the gap between them and everyone else gets wider rather than narrower. In a region like ours, that gap matters more than it does in most places.

The second is that capability building doesn’t announce well. A funded apprenticeship that puts structured AI skills into a workforce an employer already has, using their own processes and their own problems, is not a headline. It’s slower than buying licences. It’s harder to put in a press release. It also happens to be the thing that changes what people do on a Monday morning.

My hope for the next few months is fairly simple. That skills get treated as infrastructure rather than as a chapter near the back of the AI strategy. And that whoever ends up owning delivery remembers the businesses who won’t be lobbying for it.

Where this actually gets settled

Reynolds is a capable operator. Narayan understands the technology better than most people in Parliament. The cabinet seat is a good signal and I’m glad of it.

But none of it gets settled in Whitehall. It gets settled in whether a team leader in Sunderland is doing her job differently in six months’ time, and whether anybody bothered to help her get there.

Which isn’t only government’s job. It’s ours as well.

Which is why, off the back of all this, we’ve got something coming. We’ll be publishing our own proposal shortly for how we think this can actually be done across the North East, properly, with the businesses who’d usually get missed. Exciting times ahead, if I’m honest, and a bit daunting too.

If any of that sounds like something you’d want a part in, get in touch.

This was posted in Bdaily's Members' News section by We Are Zenith .