Ask any salesperson how much information they’d like on a prospective customer and you’ll never hear the end of it. That’s the premise driving the crowded sales intelligence market, which today has services that can do everything from helping to identify prospects and surface background about them, to even writing the pitch and performing autonomous follow-ups.
But sales teams need more than data; they want context. Sumble, a startup out of San Francisco, is trying to provide that context by trawling the web across social media, job boards, company sites, regulatory filings, and so on, to surface information about what’s happening inside companies.
The brainchild of Anthony Goldbloom and Ben Hamner, the founders of data science and machine learning community Kaggle, Sumble uses a knowledge graph underpinned by large language models to connect the various data points it gathers. The result, Goldbloom told TechCrunch, is a comprehensive view of a company’s technographic data — which tools are used in which departments, any projects being launched or run, its organizational chart, what tech a company might be looking to adopt, and, crucially, who to contact.
But given how crowded this market already is, from incumbents to myriad AI sales development rep agents, the question is: does the world really need more?
Goldbloom believes so, and says the startup’s approach seems to be working: He told TechCrunch that since the startup launched in April 2024, it has signed 19 enterprise customers, including Snowflake, Figma, Wiz, Vercel, and Elastic, and it has tens of thousands of users in total. About 30% of its users pay for a Pro subscription (either themselves or their company does) and so far, growth has been driven by word of mouth. The startup declined to share details of its revenue, but we understand revenue increased by 550% year-over-year.
“What tends to happen is, we go viral inside a company,” Goldbloom said. “We’ll go from 1 to 500 MAUs [monthly active users] in a company over a six-month period. And the way the spread happens is normally within a Slack channel, then within a single team, then within an office, and then within that company.”
Goldbloom said traction, the quality of customers, and strong customer retention played a big role in catching investors’ attention. The startup on Wednesday emerged from stealth with $38.5 million in funding — Coatue led an $8.5 million seed round, while Canaan Partners led a $30 million Series A. AIX Ventures, Square Peg, Bloomberg Beta, Zetta, and angel investors, including Salesforce CEO Marc Benioff and former GitHub CEO Nat Friedman, also invested.
Notably, Sumble’s co-founders have attracted investors they’re quite familiar with. Rich Boyle, currently a general partner at Canaan, was a board observer at Kaggle, while Bloomberg Beta and Zetta also were on Kaggle’s cap table. And Goldbloom co-founded AIX Ventures and is an investment partner, though he told TechCrunch he’d “stepped out of the room” when the firm considered the investment in Sumble.
Still, Sumble faces a lot of competition. Challengers include Apollo.io, Slintel, SalesLoft, Cognism, Reply.io, ZoomInfo, HubSpot, and Outreach, among others that deliver either more focused point solutions or all-in-one sales IT toolkits. And since Sumble currently uses publicly available data, there’s little to stop others from doing what it does currently.
Goldbloom, however, is confident that Sumble’s moat is deeper than it looks at first sight, thanks to how its knowledge graph is structured, covering about 2.6 million companies around the world.
“The way we think about it is, the more data that we add into the knowledge graph, the richer the corpus will be. We view the richness of the knowledge graph as a massive source of defensibility,” he said.
Sumble is also counting on the continued adoption of large language models to help it continue to scale as it expects to see people using AI along with its service. “The way we structure our data is so the knowledge graph is and always will be very query-able by large language models […] The idea being that you can ask ChatGPT about the Apple tech stack, or you can ask ChatGPT about the Apple tech stack grounded by our data,” Goldbloom said.
“We think AI is going to change the data vendor landscape a lot such that having a knowledge graph structure as a way to feed context into the large language model is going to be a key part of the LLM ecosystem,” he added.
The service is currently offered as a web app and via an API, and there’s also a paid plan that offers more features like integrations into workflows and CRMs, as well as notifications when a development at a prospect may be of interest.