A specialist magazine publisher runs two businesses that share a brand: digital platforms, web, apps, newsletters, subscriptions, where the asset is first-party audience data; and print-led workflows, commissioning, layout, production, rights and archives, where the asset is the content itself and the process that makes it. AI opportunities in the two are different in kind, and a programme that treats them as one list funds the wrong things in the wrong order. This is the split we used with one of the UK’s largest specialist publishers.

Digital: data and personalisation

The digital side is where AI already lives in most publishers, and where the next gains are:

  • Audience insight and segmentation from first-party data, which the end of third-party cookies made more valuable, not less.
  • Personalisation of content recommendation and newsletters, with the editorial team setting the boundaries.
  • Subscription growth and retention: propensity and churn modelling, offer optimisation, retention workflows.
  • Advertising and commercial operations: campaign planning, creative variants, performance reporting, sales-enablement assistants grounded in the media pack.
  • AI search and answer-engine visibility: structuring content so that AI Overviews, ChatGPT and Perplexity cite it, which is now a commercial question as much as a technical one.

The data position is usually good, because the systems of record exist, and the pathways are mostly product-led: the market has tools, and the work is selection and integration.

The print side has fewer AI products built for it and more opportunity for tailored work:

  • Editorial research and ideation, grounded in the archive and approved sources, with a person making every editorial call.
  • Consistency and house-style checking across issues and brands.
  • Layout and production automation, where repetitive composition can be drafted for a designer to finish.
  • Rights and archive mining: what the publisher owns, what it has licensed, what it can reuse, across decades of content that nobody can search today.
  • Back-office productivity: the routine work common to every function.

The archive is the key. A publisher’s back catalogue is a knowledge base nobody has indexed, and a grounded assistant over it is both an editorial tool and a rights tool. It is also usually the data-foundation project the digital opportunities do not need and the print opportunities cannot proceed without.

Where they meet

Governance. Editorial oversight is non-negotiable: AI assists research and production, people make the calls that carry the brand. IP protection cuts both ways, protecting the archive from training use and keeping generated content clear of others’ rights. Personalisation runs on first-party data under data-protection law. Any reader-facing assistant discloses that it is AI. These are design constraints from the first workshop, not a review at the end, because for a trusted media brand the trust is the product.

The roadmap shape

Two tracks with shared foundations. Digital proceeds now on product-led selections, measured on subscription and advertising metrics. Print starts with the archive and documentation foundation, then the research and production assistants on top of it. Productivity and governance run across both from day one. The result for the publisher we worked with was a clear taxonomy from low-risk productivity gains to higher-impact strategic initiatives, build-versus-buy-versus-partner guidance, and a roadmap that evolves responsibly.

Our publishing work is described under AI for publishers and media, and the Immediate Media engagement in the case studies.