It’s no fun as a customer to have an interaction with a bot when it’s clearly a bot you’re dealing with.
Rasa is a startup that claims to have developed the infrastructure to enable developers at large enterprises to build “powerful” conversational AI assistants so that those interactions feel more personal and meaningful to users. It says it does this by providing CALM (Conversational AI with Language Models) infrastructure and a low-code user interface.
Its technology has helped it attract a number of large customers, mainly in the financial services and telecommunications space. These clients include two of the world’s top three banks, two of the largest banks in the United States, American Express and Deutsche Telekom, among others.
Founded in 2016, Rasa began as an open source platform for developers to build chatbots, voice apps and other services that use conversational AI for interactivity. Since then, Rasa has been downloaded by developers more than 50 million times, the company claims.
A few years ago, the startup recognized an opportunity to help businesses increase engagement with their customers. Oracle CEO alum Melissa Gordon was hired to help guide this new strategy. Gordon, who was a pole vaulter for many years in a sport when it was not a women’s sport and used Title IX to compete on the men’s team, said she likes to “challenge the status quo.”
“We’ve always been very vocal from the beginning about challenging some of the established ideas about how chatbots should be built,” said co-founder and CTO Alan Nichol.
The move appears to have paid off for Rasa, which says it “roughly doubled” its annual recurring revenue (ARR) in 2023 compared to the previous year. That traction helped it secure $30 million in Series C funding, led by StepStone Group and PayPal Ventures, with participation from existing backers Andreessen Horowitz (a16z), Accel and Basis Set Ventures. The company declined to disclose the valuation, saying only that it was a step-up round from its $26 million Series B raise.
There’s no denying that the space is hot – and increasingly crowded -. On Tuesday, TechCrunch reported on Sierra, a conversational AI startup founded by former Salesforce co-CEO Bret Taylor and ex-Googler Clay Bavor, which claims its software can actually take actions on behalf of the customer.
Rasa is different, Nichol said, in that it is not all for agents.
What the two companies have in common is that they both claim to tackle issues like illusions, where a large language model sometimes gives an answer when it doesn’t have the information to answer accurately.
Rasa claims it is unique in that it allows businesses to “fully leverage the power of LLMs to understand the language at a really granular level, without exposing themselves to these kinds of risks.” In other words, it says it uses LLM to understand users, rather than second-guess business logic or drive a conversation, so businesses can maintain control over the flow of conversations and the things their bots say.
It’s a bold claim.
But as a result, Gordon says, the bots that Rasa’s infrastructure is helping to develop don’t feel like branded bots.
“A classic use case is a customer-facing flagship, and typically, those teams when they come to Rasa, it’s usually not their first rodeo,” Nichol said. “They were on another platform or they tried to build everything in house. And then, at some point, they’ve run out of steam, and then they’re looking for something that’s more scalable.”
Image Credits: Rasa
Bots have external use cases ranging from checking account balances to transferring money. In the case of Deutsche Telekom, for example, bots can help reset a router in someone’s home if they have an internet problem, Gordon points out. Notably, the investment marks PayPal Ventures’ first investment in artificial intelligence. In a written statement, partner Alan Du said: “At PayPal, we’ve seen how Rasa’s technology improves customer engagement and business performance through our concierge solutions, and we’re making our first AI investment in Rasa because we believe it’s the best platform for businesses to develop powerful conversational AI.”
