Our challenge to publishers: use AI to change your business like we have ours

We have spent the past four years at HBM Advisory advising publishers on how to grow their digital businesses. But over the past six months AI hasn’t just changed how we work – it has changed what we do.

At the WAN-IFRA World News Media Congress in Krakow, we told the story of how a former newspaper journalist and a former CMO became publishers again. And product people. And data science partners.

This didn’t happen because we set out to change things – we are still taking on consulting clients – but because a new generation of tools made it possible and the opportunities were too big to ignore.

We opened our talk with a quotation from David Caswell, a respected expert in AI and media. He says that he always frames the challenge of AI to editors in a positive light. He asks them “What would you do if you had 20,000 extra journalists, designers, video producers, all at almost no cost? And what would you do if every other publisher had the same?”

This is the reality AI is creating.

We spoke about how in adopting article summarisation, chatbots, transcription, translation and exploring answer engine optimisation (AEO), publishers are merely conquering the foothills of the possibilities AI is creating.

The next stage is about building systems that generate, distribute and measure journalism in new ways. We were able to use ourselves as a case study of what was possible.

Just a few months ago we started experimenting with an AI-powered search tool called NoahWire, which scans the web hourly, detects emerging stories relating to a user’s prompts and returns summaries or article drafts, complete with headline, standfirst and image.

We began feeding it prompts about areas we found interesting — trends in subscription, monetisation, new formats, AI tools — and found it was surfacing genuinely useful stories, including some from markets, such as Japan and Korea, into which we had no previous insight.

So using those feeds we launched Tomorrow’s Publisher, a daily site for publishing professionals, updated every weekday with 2–8 stories, and staffed by one person (me!). In just half an hour a day, I scan the feeds, select articles, edit them, adding new headlines, standfirsts and bashing the copy around – old editing habits die hard – and publish them.

We didn’t set out to become publishers again. But we had the tools, the ideas, a few prompts and the rest was easy.

Then we automated distribution. Using Zapier, we built a workflow that watches our RSS feed, sends new stories to ChatGPT to write a LinkedIn post (with a prompt to avoid jargon and emojis) and drops the output into a spreadsheet.

The Tomorrow’s Publisher workflow now powers newsletter and website content for the Association of Association Executives who we supply under our new brand Amplify. We also used it to build a site, SRM Today, for Suppeco, a B2B tech firm frustrated that its niche – supplier relationship management – was barely covered by the trade press. Turns out there’s a steady stream of stories there too. No subject is too obscure.

Separately, we’ve partnered with data scientists to automate the kind of content reviews we used to run manually at The Times. A review that once required eight graduates to manually categorise content over four months now takes two days with the help of AI – and it can analyse as many stories as the publisher wants.

We can combine that taxonomy with performance data to show what kind of content actually works, across the full range of a publisher’s content, not just the top 10 or bottom 10 stories, but the 95% in between too.

We ended the talk by pointing to the next mountain that needs conquering: micro-personalisation. We’re working with data scientists who have clustered the world’s adult population into 2 million personas, and are developing prompts that adapt tone, wording and calls to action to suit each one. This isn’t something that will be around “in the future”; it’s in test now.

Our intended takeaway was that if we can do this – with no engineering team, no funding round, no proprietary models – then publishers can too, especially given the resources, knowhow and incentives they have. In fact, we said, if they’re not already exploring these ideas, they are probably falling behind.

Alan Hunter is a co-founder of Tomorrow’s Publisher and HBM Advisory