.webp)
Reach has chosen Taboola's DeeperDive generative AI answer engine for several of its news sites, including the Express and Daily Star.
The tool sits on publisher websites and uses the publisher's own articles to generate responses to reader questions. It also surfaces related stories from the same site based on the query.
The deal comes as publishers assess how generative AI tools may affect audience behaviour and referral traffic. Some have argued that AI-driven search products could reduce click-throughs by presenting answers directly in search results.
On-site answers
DeeperDive uses a conversational interface that lets readers ask questions in natural language. It returns an answer and links to relevant articles from the publisher's archive and current coverage.
It can also suggest follow-up questions based on detected areas of interest. Taboola says the product is designed to provide timely, contextual responses using the publisher's reporting as source material.
Reach describes itself as the UK and Ireland's largest commercial news publisher. It publishes more than 120 brands, including national titles such as the Mirror, Express, Daily Record and Daily Star, as well as regional outlets including MyLondon, BelfastLive and the Manchester Evening News. Reach also operates the Irish Star in the US.
Reach says it reaches 70% of the UK online population each month and 10% of the US online population, and has more than 100 million social followers globally.
Product focus
DeeperDive is designed to keep interaction within a publisher's own environment rather than sending readers to an external search engine. Taboola says it draws on content produced by journalists and editors and returns answers that cite that material.
Taboola is positioning DeeperDive as a response to the rise of generative AI search tools, as publishers and platforms reassess distribution strategies and product design and more people use conversational search for information discovery.
Since launching in 2025, several publisher groups have implemented DeeperDive, Taboola says, including Gannett's USA TODAY Network, India Today Group, BuzzFeed Asia and The Bangkok Post.
Data and ads
DeeperDive uses Taboola data signals from its network of publisher partners. Taboola cites more than 600 million daily active users across 9,000 publisher partners as a source of real-time insights into what audiences are reading and engaging with.
Taboola also says the product creates opportunities for new advertising placements within AI-generated answer pages. Under the proposed model, contextually relevant adverts would appear alongside answers to questions that indicate intent, similar to search advertising, keeping advertising and user interaction on the publisher's site.
Reach Chief Product & Technology Officer Terry Hornsby said the roll-out reflects a broader shift in how audiences look for information online.
"This is an important step for us as we focus on how to reach readers where they are, with conversational search becoming increasingly common. DeeperDive gives us the ability to continue that focus, giving readers the ability to learn more and explore topics they care about. Taboola has been a great partner for us as we continue to harness generative AI and apply to our everyday user journey," Hornsby said.
Taboola Chief Executive Adam Singolda said the product changes how publisher content is presented and how readers interact with it.
"Reach PLC is taking their rich history of delivering quality content and with DeeperDive, offering it to readers in a completely new, appealing and innovative way," Singolda said. "Publishers who adopt DeeperDive aren't just keeping up with the GenAI revolution, they're leading it, on their terms. We're creating a new user habit, one where readers lean in, ask follow-up questions, and stay to explore. We're also giving publishers the potential to unlock even more ways to grow."
The deployment across Reach titles adds another major UK publisher group to the growing list of media companies experimenting with on-site generative AI features, with more expected to trial similar formats as conversational search becomes more common.