The digital publishing industry had been operating on a relatively stable, understood model for decades: users actively went to search engines like Google to look for information, where they were presented with a list of links to click through to read traditional web pages on our native sites. It was a highly imperfect but functional model with well-understood rules. However, AI is causing an evolution not only in how consumers find content, but also in more general user behaviour and expectation. The model we have operated on for decades is rapidly disappearing, and with it, publishers are being thrown into unstable, unpredictable, and unknown waters.
Customers are interacting with AI models in a more conversational way than they ever did with search engines and receiving highly personalised answers in return. In order to succeed, publishers must have a deep understanding of the varying needs of their customers at different points of interaction. Brutally put: for what use cases and customers is your brand ‘better’ than ChatGPT, and how do you maintain this advantage?
Publishers also need to be careful not to remove the opportunity for serendipitous browsing or lose sight of what makes their brand unique. The issue isn’t just that AI models tend to stick to answering within the specific personalised scope of the question, it is that users often don’t want any more than this. While this may be great for getting clear answers to defined questions, it eliminates the ability for users to stumble upon new information and discover new sources.
So, we are therefore currently caught in the contradictory nature of this threat: AI platforms need high-quality content from publishers to train their models, but publishers need direct customers to monetise in order to survive. For those of us who have navigated the digital media landscape over the last two decades, this current battle will feel eerily familiar to the early days of Google, and later Google News, when digital platforms first began pulling publisher content off native sites and into their own aggregated environments. The difference, though, is that the current situation is on a different scale: the amount of ‘zero click’ answers (search terms where no click is activated) is now approaching 60%. AI interfaces have immediately provided users with more comprehensive answers than have ever been provided before, and with less need and fewer opportunities for click through to publishers than ever before. In one fell swoop, publishers are facing the existential threat of no longer owning the direct relationship with their customers.
The front page, be it on paper or on Google, has historically told readers what they need to know, whereas an AI only answers what they want to know. So this evolution poses a critical question for publishers: in an ecosystem driven by complete, personalised AI answers, how do we lean into providing value to customers at the point of demand – often not on our own sites and apps – whilst then demonstrating that we are more than just an ‘answer engine’?
Marrying technology and journalism
To successfully navigate this new era, publishers must look for ways for technology, marketing and journalism to work in tandem, helping each other and playing to each other’s strengths to suit the user’s differing needs. It is not the case that one necessarily comes at the expense of the other; each discipline can elevate each other, with data underpinning all three.
To use Which? as an example, our most valuable asset is the trust consumers place in us. It is the trust in our brand that underpins our ability to attract users, which in turn supports our mission as a not-for-profit organisation championing UK consumers. Where does that trust come from though? Aside from 70 years of hard-earned brand heritage, it comes from top quality journalism, investigations and the millions of unique data points from our rigorous product tests and consumer research which allow consumers to choose the right products and service with absolute confidence.
How we use this data is also one of the most interesting ways in which technology and journalism can cooperate. Technology now allows us to deliver data in brilliant, interesting, useful new ways which can support our journalism. In a world where users expect responses to be hyper-personalised, bringing that detailed data to the fore to help give much more flexible, informed answers to questions on, for example, which television to buy, creates unique value to a customer, and by doing this, we can bring them into contact with our investigations, campaigns, and wider reporting.
For Which?, having excellent data has always meant we can produce the kind of insightful and informative journalism that we are famous for. We turn countless data points into a story or feature - something tangible and useful for consumers to help them make better, more informed, decisions. Now, thanks to AI, this is no longer the only way we can use data. Technological advancements have meant we can undertake a radical rethinking of how we utilise one of the best assets at our disposal.
Giving access to the raw data
It is now possible to free up this proprietary data so it can be used natively to give people better, highly personalised answers. For the first time, we can be both “data interpreters” and “data curators”. The two go hand in hand and it would certainly be a mistake to think that one comes at the expense of the other. Just because users might want personalised answers to some questions, they will equally still want all the things journalists have excelled at for years: from summaries to introductions and guides, and the personal recommendations from trusted experts that we all rely on so frequently.
To put this into action, we are developing initiatives like Ask Which?, our new AI-powered assistant designed to help consumers navigate complex product choices with confidence by allowing them to ask more specific and personalised questions about products to help them find the right choice for them.
By unleashing the raw data and giving users closer access to it than ever before, we allow users to find the products that best suit them, their needs, and their requirements. What a specific user doesn’t value can be just as ultimately important as what they do: for example, if your washing machine is tucked away in your utility room, does it matter if it runs silently? If you’re looking for a hairdryer to travel with, size and weight could be more important considerations than energy efficiency. This is even more important for life-changing decisions like what health and wellbeing interventions are worth pursuing: providing the right guidance in formats that work for customers based on their circumstances is so much more powerful than generic advice. While we as publishers factor in all these (usually important) variables, as data curators, we allow users to choose their own priorities, confident that they can trust the information they are receiving.
But this only works if you are comfortable with your jumping off point. To know what questions to ask, or even to be comfortable and confident in what features matter to you, you need to know about what you’re dealing with. In that case, the stories and guides produced by journalists can provide the much-needed introductions, setting the scene that then allows users to explore subjects in more personalised ways.
Beyond the product itself, both content and AI access must be placed before or after paywalls highly strategically. The process of getting users to register must be much more creative than just gating the best content behind a traditional barrier. It should include exploring models that offer free direct access to draw users into a trusted ecosystem, with paid-for access offering genuine interactive value, such as making content “saveable” and useful over time. The entire proposition should be framed as a fair exchange of value for user data. Free registered access becomes a critical tool in giving customers the option to explore and understand value and at Which?, we are constantly experimenting with what the value proposition in this free layer should look like.
The landscape continues to evolve at a fast pace, and every change throws up more things for us to consider – be they opportunities or potential threats. Previously, the industry operated under a relatively straightforward dynamic with Google almost exclusively mediating customers for the entire news industry. As we are now operating in a fragmented, multi-platform landscape, publishers are no longer beholden to a single monopoly for customer mediation – something that we would have been thrilled about 10 or 15 years ago. A diversified platform ecosystem offers publishers more leverage and varied avenues for negotiation, variation in strategy, trial and error.
The future of digital publishing lies in a harmonious partnership where journalism and technology work hand in hand. While AI platforms present significant challenges, technology also provides the tools to help users discover high-quality, rigorous journalism in an increasingly automated, personalised landscape. In a world saturated with AI-generated content, consumer trust remains the essential antidote, ensuring that human expertise and verified data continue to guide and empower audiences. Looking for ways we can marry technology and journalism will give publishers the opportunity to forge a profitable and exciting path forward in this new technological era, that not only gets their content in front of users, but also allows users to interact with their content in the way they want.
This article was first published in InPublishing magazine. If you would like to be added to the free mailing list to receive the magazine, please register here.
