How HBM Advisory Publishes an Automated Media Industry Website and Newsletter

Tomorrow’s Publisher is an automated news website. But it is everything that the ever-increasing AI slop all around us is not. It doesn’t plagiarize or fabricate “news” out of thin air. It is not plastered with bottom-feeder ads. TP offers timely, well curated and relevant news for the media industry from trusted sources with minimal but still crucial editorial oversight. And last but not least: It has a sleek design.

TP’s news production model challenges today’s prevailing assumption that AI should be used primarily in the backend for optimizing subscriber journeys and SEO, for translating and re-formatting existing content, for chatbots or for research, but definitely not for publishing a whole journalistic website. (When the Italian news site Il Foglio tried that a few months ago, the publisher deemed the experiment a commercial success, but that was most likely a result of the widespread international coverage of the weird journalism “Foglio AI” produced.)

During the first Chefrunde Study Tour (Crusa) in London last week, which I co-organized, our group met with Michael Brunt and Alan Hunter, two journalists turned media consultants. They are the team behind Tomorrow’s Publisher. Before co-founding the media consultancy HBM Advisory four years ago, Alan Hunter was Head of Digital at The Times [of London] and Michael Brunt was General Manager at The Times and The Sunday Times.

During our meeting, they explained how Tomorrow’s Publisher works, why they started it as an experiment, where they are taking it and how it fits into the bigger picture of how AI is shaping the news industry.

This analysis is based on our meeting in London plus their answers to a few additional questions I later asked them via email.

Alan Hunter (background left with laptop) and Michael Brunt (right with laptop) discussing Tomorrow’s Publisher and the wider industry context with the Crusa group

TP produces 2-8 daily stories with one staff member working on it only 30 minutes per day, using the AI discovery engine Noah to scan 500,000 sources and generate publication-ready content faster than traditional trade press.

As Alan Hunter explains: "We are able to jump on stories really quickly. From it first appearing in my feed to publishing it can take just a few minutes.”

The system combines three components:

Noah's verification process provides built-in safeguards: "Once it spots a potential story that meets our prompts, it will crawl the web seeking confirmation of the story (where possible) and the facts therein," explains Hunter. This allows sourcing beyond mainstream publications while maintaining accuracy. Publicly available information can be systematically harvested and processed faster than by using traditional newsroom workflows. However, HBM acknowledges the advantage of operating in publishing industry news, where "there is rarely much controversial news, let alone disinformation”.

TP began as an experiment to see how Noah's technology could be used and whether it could produce a publication that would engage people. TP attracts a solid, and growing, four-figure audience per month and its weekly newsletter goes out to more than 500 recipients. “We are aware these are not earth-shattering numbers but we are confident we are reaching the right people”, says Hunter.

On top of this experiment HBM built a business model when they realized "we could do this for other people!" So they began offering the same AI-assisted publishing service via their product Amplify to clients who want content for their websites or newsletters.

Amplify creates specialized publications for very narrow industry verticals, using the same AI technology (Noah + human editing) that powers Tomorrow's Publisher. Using Amplify, publishers can serve markets that are too specialized to cover profitably the traditional (human-powered) way.

For example SRM Today: this vertical covers supplier relationship management and demonstrates AI's capacity to serve previously uneconomical niches and attract industry advertising. Also among their clients are an audiology clinic, a software provider and a number of trade associations.

The model scales through "fractional editors”. Instead of hiring full-time editors for each niche publication, they use part-time specialists who work a fraction of normal hours.

Each editor specializes in a specific vertical (like supplier relationship management, audiology, etc.) AI does the heavy lifting (finding stories, writing first drafts), so human editors only need to spend minimal time on curation and quality control.

The use of AI makes the delivery of “unsexy” but very relevant content for certain niche target groups economically viable. This is similar to Morning Brew’s B2B business model, but HBM has more fully integrated AI into the whole workflow.

Excerpts of HBM Advisory’s value proposition, as presented to the Crusa delegates

HBM also uses AI to figure out what makes content work. They systematically analyze articles across different factors:

This analysis shows why specific content performs well or poorly. Insights traditionally lost in subjective editorial judgment become visible. Publishers can optimize content strategy based on empirical patterns rather than intuition. The data confirms insights that Hunter already discovered through trial and error in his days at The Times.

Three key takeaways emerge from HBM's approach:

Volume Reduction Works: HBM's experience confirms that reducing article count by 30% increased engagement, contradicting the "more content" response to declining traffic. As Alan Hunter notes: "At The Times [of London], I first cut the number of articles on the website by 10% and nobody noticed. Then by 20% and still nobody noticed. And finally by 30% and the only people that noticed were the reporters. And the engagement went up. Many publishers have reported the same effect of reducing their content." Even loyal readers consume only 3-4 articles daily.

Niche Markets Become Viable: AI economics make previously unserviceable specialized topics commercially feasible. Every industry vertical becomes a potential publishing opportunity.

Personalization at Scale: HBM has a vision of "two million personas”. Each individual reader gets content tailored to their specific preferences. It’s still the same story, but presented differently for different people. One person might get bullet points, another gets long-form narrative. Headlines could be data-driven for some, emotional for others. And content depth varies based on expertise level. AI can automatically adapt the same core information into formats that match individual reading patterns, interests, and consumption habits. Think of it like having two million different editors, each crafting content for one specific person.

This is still largely theoretical. HBM mentions it as "the next stage" and shows it as a future vision rather than something they're currently doing. The technology exists to personalize at scale, but most publishers haven't implemented this level of individualization yet. It represents the ultimate endpoint of AI-driven publishing: mass customization of content for individual consumption preferences.

Publishers face a brutal reality: traffic down 50%+, falling ad rates, and readers increasingly turning to ChatGPT for quick answers. As Alan Hunter observes: "Privately, people describe traffic falls that are way beyond what's already been reported. A lot of people are saying publicly that they've lost 50% of their traffic in the past two years. That's not really the effect of AI - not yet anyway - but Google algorithm changes."

The old model of high-volume, low-value content is breaking down. Publishers who survive are becoming trusted "destinations" where readers go for analysis they can't get elsewhere. Michael Brunt emphasizes this shift: "Publications are imagining themselves as more of a destination, particularly those that have the brand trust and integrity. If a big event has happened - for example, the recent conflict between Iran and Israel - I will go to The Hindu or I will go to The Times to identify what it means for me, because I know what they stand for."

HBM's model bridges the gap between commodity and exclusivity. Tomorrow's Publisher isn't just aggregating press releases - it's curating and contextualizing industry developments for professionals who need expert filtering. The AI handles the discovery and initial processing, but human editorial judgment ensures the content serves a specific community's needs. This creates a middle path: faster than traditional journalism, but more valuable than raw information feeds.

The economics work because AI dramatically lowers production costs while maintaining editorial standards. Publishers can serve specialized audiences profitably, even with modest subscriber numbers.

HBM Advisory's practical demonstration proves AI can automate significant portions of information publishing while maintaining editorial standards. Their 30-minute daily workflow producing professional publication output represents a new baseline for publishing efficiency.

The broader implication: information abundance makes curation and analysis just as valuable as raw reporting. Publishers who recognize this shift and implement appropriate AI tools maintain a competitive advantage in an increasingly automated landscape.

Tools used for this newsletter: