Meta Acquires AI Analytics Start-Up Sundial

Meta has acquired Sundial, a start-up whose software lets staff ask questions of company data in plain language, bringing its two co-founders, both former Meta employees, back to the company. Neither side has disclosed the terms.

“Today, we’re excited to announce that Sundial has been acquired by Meta!” co-founder Chandra Narayanan wrote in a LinkedIn post on 23 October. A Meta spokesperson confirmed the deal to Axios, which first reported it.

Narayanan will lead Meta’s analytics tools and agents work

Narayanan says he is rejoining Meta “along with Julie Zhuo, Stanley Watt, Asa Levi, John Maier, and other members of our incredible team, to build the next generation of AI analytics tools and agents”. He left the company nine years ago, he writes, and the move “feels like coming home”.

Both founders have long Meta histories. According to Axios, Zhuo was previously Meta’s vice president of product design, and Narayanan was director of analytics at Facebook before leading analytics at Instagram and Sequoia Capital. Other Sundial staff, including several Meta alumni, are joining too.

Sundial will be folded into Meta’s new data analytics organisation, led by chief data officer Alex Schultz, Axios reports, citing a LinkedIn post by Schultz. Narayanan will run the group’s “data analytics tooling and agents work”, Schultz said, according to The Verge.

“AI has created effectively unlimited demand for insights, and we want to define the future of AI analytics,” Schultz wrote, as quoted by Axios.

Sundial turned plain-language questions into data analysis

Axios describes Sundial’s platform as one that turns natural-language questions and complex data into simple insights through a chatbot plug-in. The idea is that teams can analyse data themselves instead of waiting for a central data team to respond.

The product is not expected to stay on the market. Axios reports that Sundial will stop serving external customers when the deal closes.

In his post, Narayanan thanked investors “Alfred Lin at Sequoia Capital, Mike Vernal at Conviction, and David Fischer at 01A Advisors”, and singled out “David Sasaki and his team at OpenAI, who pushed us harder and further than almost anyone”.

Meta made analytics a board-level job in July

The deal builds on a reorganisation earlier this year. Meta named Schultz, its marketing chief, as its first chief data officer in July. Meta’s leadership page says he oversees analytics, data science, data engineering, competitive intelligence and user research, and leads “Meta’s transition to AI-led analytics and research”.

According to Axios, his unit is called Data Analytics Tooling & Agents, or DATA. It handles descriptive analysis, including finding questions the company has not yet asked, runs experiments to improve Meta’s data and AI products, and monitors critical AI systems.

When he took the job, Schultz wrote in a LinkedIn post that AI was creating demand for insights “while the traditional ways we generate those insights are reaching their limits”. He said Meta’s Analytics Agent was already “the most widely used AI agent inside Meta”. In an interview with Axios in July, he named a “semantic” layer across Meta’s data warehouse as a top priority.

The deal comes days after Meta’s third-quarter results, in which it said it would create a new class of non-voting shares partly to fund “potential acquisitions”, and on the same day it took its Muse AI agent to Japan. Elsewhere in AI this month, Prime Intellect bought Moondream with its founders taking charge of specific work.

Meta is buying a team to point AI at its own data

Meta has not given Sundial’s price, but its interest is in the people and the product rather than the customers: by Axios’s account the service will stop selling to outside companies, and its founders will run the very tools Schultz has called the centre of how Meta learns.

That makes it closer to the Moondream deal, where founders came in to lead specific work, than to a bid for market share. The appeal for Meta is speed. A team that has already built a plain-language analytics product, and that knows Meta’s culture from the inside, can be pointed at Meta’s own data rather than built from scratch.

Plenty is still unknown. Neither company has said what it paid, when the deal closes or how Sundial’s software will fit alongside Analytics Agent, and the DATA unit’s semantic-layer work is internal and hard to judge from outside. What is on the record is the direction: Meta now has a chief data officer, a unit named for analytics agents and, from Sundial, a leader for it, nine years after he left.