Financial Times leads in AI-Driven hybrid search with MongoDB

OUR SOLUTION

Harnessing expertise to develop serendipitous content discovery in just 18 weeks

The existing relationship meant the team was already MongoDB proficient, so no new learning was required. Additionally, the search technology MongoDB Atlas provided, including MongoDB Atlas Vector Search, would enable the FT to efficiently deploy full-text search and vector search capabilities. It also enabled the FT to combine both capabilities for unified results in a hybrid search. Sealing the deal was the wraparound support and expertise provided by the MongoDB professional services team—a crucial factor given that the FT is an early adopter of hybrid search. “It’s important to know that you can rely on good support,” said Terziev. “Search is a complex space. It encompasses areas such as information retrieval, statistical analysis, AI, [machine learning], and user experiences. Some areas we felt confident with; in others we wanted external expertise.”

Using machine learning, vector search converts content into high-dimensional numerical vectors and measures their proximity—bringing the most relevant pieces to the forefront. By blending traditional keyword search with the context-aware intelligence of vector search, hybrid search delivers recommendations that are both precise and meaningful. In so doing, it fulfils the FT’s ambition for more serendipitous content discovery. “Probably the key challenge for any platform is being able to predict the users’ needs before they’re articulated,” said Terziev.

The initial deployment milestone, releasing hybrid search for web users on FT.com, was completed in approximately 18 weeks, going live at the end of September 2024. “We gained a lot through being able to engage [with] MongoDB professional services,” said Terziev. “[It was] very, very helpful, especially at the later stage of the project, which is the critical part in terms of tinkering.”

THE OUTCOME

Enhancing search for busy readers and positioning the Financial Times as a media leader

With full-text search and vector search capabilities running directly on MongoDB Atlas, the FT now delivers hyper-relevant recommendations without additional search infrastructure. The deployment has positioned the FT as a leader in hybrid search within the media sector, offering readers highly relevant article suggestions—without the need for manual searches: a boon for time-poor subscribers.

Next, the hybrid search rollout will extend to the FT’s mobile apps, followed by its specialist titles, ensuring all users will benefit from its enhanced capabilities. Ultimately, the goal is to make content across all FT platforms accessible through a unified hybrid search system.

“We are at the beginning, and we expect to improve it further,” said Terziev. “It’s a journey; in certain cases, we might consider that there is no end destination because the search needs evolve as users’ demands evolve, especially now we are living in this AI-enabled world.”

“There is no one magical configuration you can apply across your search indexes and everything will be working,” added Terziev. “Building hybrid search requires domain expertise, understanding the content that’s been searched and the use case that you are optimizing and building for.” Throughout the journey, the FT’s focus will remain on surfacing content that is of utmost relevance to its readers in the most efficient way. 

What is certain is the partnership that endures with MongoDB. “What drew us to MongoDB is that it’s a full package,” said Terziev. “You have a very easy-to-use console—people who are not very technical can work with it, and it abstracts a lot of the operational work. MongoDB Atlas isn’t just a database—it’s a ‘batteries included’ solution.”