Home / Business / Facebook is buying UK’s Bloomsbury AI to ramp up natural language tech in London

Facebook is buying UK’s Bloomsbury AI to ramp up natural language tech in London

Perhaps rightly, there has long been a perception that Google-owned Deepmind has been the most aggressive in hoovering up a lot of the U.K.’s best talent in artificial intelligence, but now Facebook appears to be turning its eye to the country.

TechCrunch understands that the social network behemoth is acquiring London-based Bloomsbury AI, a startup that has built natural language processing (NLP) technology to help machines answer questions based on information gleaned from documents. According to sources, Facebook plans to deploy the company’s team and tech to work on combatting fake news and to tackle other content issues.

Bloomsbury is an alumni of Entrepreneur First. the company builder that invests in technical and domain expertise talent and helps those individuals start companies. The startup is also backed by Fly.VC, Seedcamp, IQ Capital, UCL Technology Fund, and the U.K. tax payer-funded London Co-investment Fund.

William Tunstall-Pedoe, who was instrumental in the development of Amazon’s AI-powered digital assistant Alexa, is also an angel investor in Bloomsbury.

Multiple sources say Facebook is paying between $23 million and $30 million to acquire Bloomsbury AI, in a deal that will see a mixture of cash and stock change hands. In one scenario, the startup’s investors will receive around $5.5 million, with Bloomsbury’s founding team in line for the remaining $17.5 million, paid in restricted Facebook stock. Either way, this represents a modest return for the bulk of investors, although EF — given that it invests pre-seed — is likely to have had a larger multiple.

Given the price and the stage Bloomsbury AI were at, the acquisition also has more than a whiff of acqui-hire to it, although there is some IP in the deal. I understand from one source that Bloomsbury AI’s CTO/Head of Research, Sebastian Riedel, was the biggest draw. He is considered to be a leading expert in the area of NLP, and is a professor at UCL. According to his LinkedIn, he also co-founded and is an advisor to Factmata, the U.K. startup that purports to have developed tools to help brands combat “fake news”.

Which brings us to the possible reason for why Facebook is acquiring Bloomsbury AI, a startup that I’m told was phenomenally strong when viewed as a group of researchers, but less so when it comes to getting a commercially viable product out of the door. The company’s sole product is an API called Cape that lets developers add question & answer functionality to websites and other documents.

Indeed, a source who claims to have some knowledge of Facebook’s intentions says the U.S. tech giant may be planning to put the Bloomsbury AI team on the task of helping it develop technology to fight fake news on the platform and solve other aspects of its glaring moderation problem.

Other areas of Facebook’s product that might benefit from the Q&A technology that powers Cape include being used as a workplace tool for companies to discover content in documents, or on Facebook’s consumer offering as a way of significantly improving its search and knowledge-base functionality.

It is also understood that Bloomsbury AI being based in London was a factor, as Facebook aims to have an AI presence in the U.K. capital city and is thought to be sourcing further acquisitions here.

Multiple sources have confirmed the deal to us although Facebook has not responded yet to our request for comment.

Additional reporting by Ingrid Lunden

Check Also

Datacoral raises $10M Series A for its data infrastructure service

Datacoral aims to make it easier for enterprises to build data products by abstracting away all of the complex infrastructure to organize and process data. The company today announced that it has raised a $10 million Series A financing round led by Madrona Venture Group, with participation from Social Captial, which also led its $4 million seed round in 2017. Datacoral CEO Raghu Murthy tells me that the company plans to use the new funding to grow its business team in order to be able to reach more potential customers and to expand its engineering team as well. The promise of Datacoral is to offer enterprises an end-to-end data infrastructure that will allow businesses and their data scientists to focus on generating insights over having to manage and integrate their data sources. Since nobody wants to move large amounts of data between clouds — and take the performance hit that comes with that — Datacoral sits right inside a company’s AWS systems. It’s still a fully managed service, though, but the data is encrypted and never leaves a customer’s virtual private cloud. “As companies look to their data to deliver value – data practitioners are finding that configuring and managing their own data infrastructure is a time-consuming job that is expensive and fraught with errors,” said Murthy. “We have built a platform that easily and automatically brings together data from different applications and databases, organizes that data in any query engine and acts on insights that are critical to running their business. A crucial component is that it works securely and privately within the customer’s cloud, instead of us ingesting data from their systems.” Murthy was an early engineer at Facebook and part of the team that was in charge of scaling that company’s data infrastructure and ran a part of the engineering team at Bebop, Diane Greene’s startup that was later acquired by Google. To scale Datacoral, the team is betting on a serverless platform itself. It’s making extensive use of AWS Lambda and other PaaS solutions on Amazon’s cloud computing platform. That doesn’t mean Datacoral plans to only support AWS, though. Murthy tells me that Azure support is next. “We plan to work across all of the top cloud providers by leveraging their unique services and provide a consistent ‘data-centric interface’ to our customers — essentially be ‘cloud best’ instead of ‘cloud agnostic.'” Current Datacoral users include Greenhouse, Front, Ezetap, Swing Education, mPharma and Mason Finance.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Disclaimer: Trading in bitcoins or other digital currencies carries a high level of risk and can result in the total loss of the invested capital. theonlinetech.org does not provide investment advice, but only reflects its own opinion. Please ensure that if you trade or invest in bitcoins or other digital currencies (for example, investing in cloud mining services) you fully understand the risks involved! Please also note that some external links are affiliate links.