Gannett, the publisher behind USA Today and 220 other publications, is rolling out DeeperDive, a chat-style tool that talks with readers, condenses reporting into brief takeaways, and points to related stories from across its sites. The company says the feature will sit on the USA Today homepage and replace the standard search box, offering suggested questions and short, sourced answers alongside links to full articles.
“Visitors now have a trusted AI answer engine on our platform for anything they want to engage with, anything they want to ask,” said Mike Reed, CEO of Gannett and the USA Today Network, at an AI summit in New York, an event that brought together voices from the tech industry, politics, and the world of media, “and it is performing really great.”
The move comes as publishers confront a difficult shift: AI chat systems trained on news content can create summaries that reduce the number of visits news sites used to get from search services. Reed warned that Google’s AI Overview feature has cut traffic across the industry. “We are watching the same movie as everyone else is watching,” Reed said ahead of today’s announcement. “We can see some risk in the future to any content distribution model that is based primarily on SEO optimization.”
Gannett has already made licensing deals with some AI vendors, including Amazon and Perplexity, and takes steps to block automated scrapers that crawl sites to copy material. DeeperDive represents a wager that applying generative artificial intelligence on the publisher’s own terms can bring readers back by giving them immediate answers tied directly to the newsroom’s reporting.
The widget suggests likely questions for readers to pose and offers a short synthesized reply paired with links to relevant coverage from across the USA Today network. One of the prompts shown on rollout is, “How does Trump’s Fed policy affect the economy?” When a reader selects a prompt or types a query, the tool returns a compact explanation and a set of stories that expand on the points it raises.
Reed stresses that the system is intended to rely on verified reporting rather than commentary. “We only look at our real journalism,” he says, noting the team has limited the tool’s sources to reporting it considers factual. He added that the interaction data may reveal what topics draw audience interest, a metric that can feed into commercial planning. “That can help us from a revenue standpoint,” he said.
DeeperDive was built by Taboola, the advertising company. Adam Singola, Taboola’s CEO, says the firm created the product by fine-tuning multiple open source models and baking in newsroom sourcing. Singola notes the tool draws on data tied to Taboola’s reach across publishers, describing that footprint as more than 600 million daily readers across around 11,000 publishers. He adds that the system “grounds every answer in articles retrieved from our publisher partners and requires sentence-level citations to those sources” and is designed not to produce a response when reporting from separate sources appears to conflict.
Reed told attendees that Gannett plans to test features aimed at commerce and purchase decisions, working with Taboola to build what he called agentic tools that can help readers shop. “Our audiences have a higher intent to purchase to begin with,” he says. “That's really the next step here.”
The product launch reflects a wider publisher strategy: pair proprietary reporting with AI tools that keep readers on brand-owned pages while offering quick, sourced answers that funnel readers into full articles. The company is positioning DeeperDive as a way to surface journalism, capture attention, and collect signals about what readers want without relying solely on search-driven traffic flows.