Key takeaways
| Insight | Why it matters |
|---|---|
| AI should remove repetitive administration. | Publishers can give journalists more time for reporting, editing and original thinking when technology handles routine tasks. |
| Product and editorial teams need to work together. | Subscription, data and audience products work better when they support clear editorial goals rather than operate separately. |
| Ethics must form part of implementation. | Clear frameworks help organisations assess risk, accountability, bias and the effect of automated systems on readers and staff. |
| Inclusion improves product decisions. | Teams with varied experiences can identify gaps in websites, AI systems and audience services before those gaps affect readers. |
| Technology needs a human voice. | Automation can improve pace, but trust still depends on the judgement, tone and responsibility of journalists and editors. |
| Modernisation should connect to audience growth. | Data platforms and subscription improvements should support useful products, stronger engagement and sustainable revenue. |
Artificial intelligence has moved from a specialist discussion into the daily planning of publishing companies. Editorial teams are testing research assistants, product departments are reviewing data platforms and commercial leaders are assessing new subscription journeys. Yet the most useful question is not how quickly a publisher can adopt AI. It is how technology can improve the work without removing the judgement that makes journalism valuable.
A recent interview published by FIPP provides a useful case study. Heather Payne, Chief Technology and Product Officer at Immediate Media, describes a role that joins digital development, audience growth, subscription modernisation and responsible technology. Her comments offer lessons for publishing leaders that want practical progress rather than technology for its own sake.
This article examines four themes from that discussion: reducing manual work, building responsible systems, designing for wider audiences and connecting product development to sustainable publishing. Each theme gives publishers a clear starting point for their own plans.
AI in publishing should support editorial judgement
Remove routine work, not responsibility
The strongest case for AI in publishing is often found in the tasks that consume time without creating much distinctive value. A system might help organise information, identify patterns in reader behaviour, prepare a first transcription or support routine production work. That can give journalists and editors more time for interviews, fact checking, analysis and storytelling.
However, the division of labour needs careful thought. Technology can assist with a process, but a named member of staff should remain responsible for the decision that affects the published result. That principle matters for headlines, summaries, recommendations, moderation and any material that could mislead a reader.
Payne described the aim as helping teams work faster while keeping people as the voices and faces of their brands. That balance offers a useful test for every proposed AI project. Will the system remove friction from a sound editorial process, or will it make it harder for staff to understand how a decision was reached?
A practical publishing workflow could include the following stages:
- Define the editorial task and the problem that technology should solve.
- Set limits on what the system can create, change or recommend.
- Assign human ownership for review, accuracy and final approval.
- Record errors and reader feedback so the process improves over time.
- Review whether the project saves time without reducing quality or trust.
This approach keeps AI in a supporting role. It also gives staff a clear way to challenge an output when the system produces an incomplete, biased or inaccurate result.
Publishing product teams need responsible technology frameworks
Turn principles into working rules
Ethical technology cannot remain a statement on a website. Publishers need working rules that guide procurement, testing, deployment and review. Those rules should cover data protection, security, transparency, bias, accessibility and the right level of human oversight.
The interview highlights the value of connecting business practice with academic thinking about responsible technology. That connection can help a publisher ask better questions before a product reaches readers. For example, does an AI feature work equally well for different groups? Can staff explain its output? Does it use data for a purpose the reader would reasonably expect?
Publishers should also separate experimentation from public deployment. A small internal test can reveal weaknesses without exposing the full readership to an unproven feature. Once a project moves towards launch, the organisation should set measurable standards for accuracy, performance and complaints handling.
Leadership teams can strengthen this process by creating a short assessment for every significant AI or data project. The assessment might include:
- The purpose of the feature and the audience problem it addresses.
- The data used, its source, its retention period and access controls.
- The risks of error, exclusion, manipulation or unintended disclosure.
- The staff member responsible for monitoring and escalation.
- The evidence required before the feature expands to more readers.
These checks do not need to stop progress. Instead, they make progress easier to defend. They give editors, product managers and senior leaders a common language for discussing risk, which is especially important when an organisation is managing several experiments at once.
Publishers can also consult the Reuters Institute Digital News Report when assessing changing audience expectations around trust, discovery and digital news. External research can support internal decisions, although each publisher still needs to test its own audience and editorial context.
Inclusive publishing products reach more readers
Design with different experiences in mind
Inclusion is not a separate project that begins after a product has been built. It affects recruitment, research, testing, language, accessibility and the assumptions that shape a reader journey. Payne connects her own experiences with her focus on making technology more accessible and ensuring that different voices have a place in decision making.
That lesson applies across publishing. A subscription form may work well for a confident desktop user but create barriers for someone using assistive technology. An automated recommendation system may perform well for a large audience while repeatedly ignoring smaller groups. A content workflow may favour one working style and make it harder for neurodiverse staff to contribute effectively.
Teams can find these problems earlier by widening user research and staff feedback. This does not mean asking every person to represent a whole community. It means creating enough variety in research and testing to challenge assumptions before they become embedded in the product.
Useful actions include:
- Test websites and subscription journeys with people who use assistive technology.
- Review AI outputs for gender, cultural, language and accessibility bias.
- Invite staff from editorial, product, commercial and support teams into testing.
- Measure completion rates and complaints across different audience groups.
- Explain how readers can report a problem and receive a response.
Inclusive design also improves commercial thinking. A product that works for more people can support stronger engagement, better retention and a clearer understanding of audience needs. The result is not only a fairer service. It is often a more useful service for everyone.
Publishers should make inclusion part of the product brief from the beginning. When it appears only as a final check, teams may lack the time or budget to fix a fundamental design choice. Early involvement gives product leaders more options and helps editorial teams understand how the service will feel to different readers.
Digital transformation must connect to audience value
Modern platforms need a clear publishing purpose
Modernising a data platform or subscription journey can sound technical, but the real measure is the value created for readers and the organisation. A faster system matters because it may help a reader find relevant journalism, allow a publisher to understand engagement or reduce the effort required to manage a subscription.
Immediate Media’s focus on digital capabilities, subscriptions and products reflects a wider shift across the sector. Publishers are no longer treating technology as an invisible back-office concern. Product decisions influence how audiences discover stories, register, subscribe, return and recommend content to others.
That connection requires close cooperation. Editorial teams understand the purpose and standards of the journalism. Product teams understand user journeys and testing. Data specialists can identify patterns, while commercial teams assess whether an idea supports sustainable revenue. When these groups share a clear objective, technology becomes part of the publishing strategy rather than a disconnected project.
A sensible roadmap should identify:
- The audience need that the project will address.
- The editorial or commercial outcome that will show progress.
- The data required and the safeguards that will apply.
- The smallest useful test that can produce evidence.
- The decision point for expanding, changing or stopping the work.
Publishers can track practical measures such as subscription completion, returning visits, time saved in production and the quality of reader feedback. They should avoid treating activity alone as success. More automation, more features or more collected data do not automatically create a better publishing business.
For organisations reviewing their technology stack, the World Association of News Publishers provides industry research and professional resources. Publishers can also explore the Publishrs publishing platform when they want to assess how a focused publishing system could support editorial workflow and digital growth.
The central lesson is straightforward. A platform should make it easier to deliver trusted journalism, understand audiences and develop useful products. If it cannot do one of those things, its place in the roadmap deserves fresh scrutiny.
Frequently asked questions
How should a publisher begin using AI?
Start with a defined internal task that creates unnecessary manual work, such as transcription or information organisation. Set a human review process, measure the results and expand only when the system meets clear editorial and data standards.
Can AI write publishable journalism without editorial review?
AI output still needs human review for accuracy, context, tone, copyright and fairness. A publisher should assign responsibility for every published item and make sure staff can challenge or correct an automated result.
Why do publishers need an AI ethics framework?
A framework helps teams make consistent decisions about data, bias, transparency, security and accountability. It also gives senior leaders a way to assess experiments before they affect readers or the reputation of a trusted brand.
How does inclusion affect publishing technology?
Inclusive research helps publishers identify barriers in websites, subscription journeys and automated services. It can improve accessibility and audience reach while reducing the risk that a product works well only for its largest or most typical users.
What should publishers measure after launching an AI feature?
Measures should include accuracy, time saved, user satisfaction, complaints, accessibility and any differences in performance between audience groups. Commercial results matter, but they should sit alongside editorial quality and reader trust.
How can editorial and product teams collaborate better?
Both teams should agree on the audience need, the editorial purpose and the evidence required for success. Regular testing sessions and shared ownership reduce the risk that a technical feature develops without a clear publishing benefit.
Publishers do not need to choose between human journalism and useful technology. The stronger route combines responsible AI, inclusive product design and clear editorial ownership. Organisations that want to review their publishing workflow can explore a conversation with the Publishrs team about practical digital publishing needs.
This article provides general information about publishing industry trends and best practices. For specific advice about implementing new systems or processes at your publication, we recommend consulting with your technical and editorial teams.
