What’s your AI vision?

Fresh thinking at Twipe’s Stuttgart summit

I arrived in Stuttgart earlier this week feeling exhausted after the worst night’s sleep I’ve had this decade – a combination of too much coffee, rich food and too many miles on the clock. But I left it positively energised after attending Twipe’s Digital Growth Summit.

The subheading for the conference was “AI, tech and the future of news” and for once the tone at a media event was upbeat, not morose, and forward-facing, not reflexively backward-looking.

Here are a few of my takeaways (Sarah Cool-Fergus of Twipe has posted hers here):

1 Publishers need an AI vision

News organisations have always been very good at defining their mission (why they do things), but not so good at setting out their vision (where they want to get to). This must change.

As I found out, people in the industry are crying out for such direction.

I ran a workshop at the summit entitled “From experiments to enterprise: how to scale AI use in your news business”. I told attendees that we would structure it as a mini-consultancy project, with a few crucial differences from what Michael and I do at HBM Advisory, notably that they would do all the work themselves and it would only last an hour (could this scale, I wondered…).

We raced through what they do with AI now; what they wanted to do; what was stopping them doing those things; and how they could break down the obstacles.

The final section was telling: three of the five groups we had divided into said right away that they needed a vision from their organisation’s leadership about where they wanted to go with AI. This would, they said, focus efforts and serve to unlock the resources of the organisation for AI projects. It could hardly be a clearer request.

2 Ask yourself: why?

I have heard Nikita Roy, host of the Newsroom Robots podcast, speak twice and each time been impressed with the clarity with which she expresses the issues surrounding journalism in the world of AI.

This week she spoke among other things about how the article is no longer the base unit of journalism because “AI does not see a story but a dataset”; how the UX of journalism is changing rapidly and becoming a conversation rather than a reading experience; and how, if you believe that services like ChatGPT Pulse will break through with users, then “the future of news is personalised, proactive and predictive”.

But most telling was her call for those present to adopt an Aristotelian approach to addressing which path to take. To do that you have to go back to first principles about what you are offering to people and why, and then build from there.

She said: “If you knew nothing about newsrooms, only that people need trusted, verified information, what would you build with today’s technology?” It’s a great question that we should all be asking ourselves.

3 The public are getting used to AI

For a cold-eyed look at the reality on the ground, you can rely on Nic Newman of the Reuters Institute to have a facts-not-vibes-based answer. This week he was drawing on the research done for the institute’s latest study, Generative AI and News Report 2025.

My takeaway is that while he said that people were “uncomfortable” with the use of AI in the writing of news, they are using it more and more. For a start, the number of people using AI to access news weekly has doubled in a year, albeit at a low level – from 3% to 6%. Young people, notably, are using it to “navigate or remix the news, to make it easier to understand”, while older people just want to get their news in the same way as ever.

The most telling number for me was that half of people trust AI search answers. For all they are a bit uncomfortable, they are getting used to AI engines and believing what they tell them.

4 You need to build an AI-first org

As I found out in my workshop, everyone is experimenting with AI. But as Gregor Zoller, director of IT at German regional publishers MHS, said: “AI usage needs to become AI thinking.”

It was an observation echoed throughout the day. One participant, speaking under Chatham House rules, said: “We are in a ‘plus AI’ world – we need to change to an ‘AI first’ world.”

Dylan Jacques, Chief Product & Technology Officer at The Telegraph, described the work his teams have been doing with AI as “building a muscle”. Which clearly you only do if you want to do heavy lifting later on.

This shift is coming very soon. If you’re not reimagining how your organisation looks in this environment then you should be (we can help with this – if you interested message me at alan@hbmadvisory.com).

5 The next decade is about agents

In some parts, the shift is happening now. Markus Franz, CTO at Ippen Digital, talked about how with a team of five he is working on “transformational innovation” powered by agents. The hope is that this will become “disruptive innovation”.

It feels pretty like that already. Among other things, they have built agents to find the next best watches on Netflix. Another set then creates articles and YouTube videos to outline the choices and posts them on a portal conceived and executed by agents.

The work is “heavily bounded by human instructions” but still the possibilities of this kind of automated workflow for publishers big and small (particularly small) seem boundless.

I think I’ll be spending the next few months researching agents. I suggest you do too.