AI startup Decagon, which is building customer service-focused agents, is in talks to raise $100 million in fresh funding at a $1.5 billion valuation, according to five sources familiar with the deal.
Andreessen Horowitz and Accel are leading the round, with participation from existing investors, according to four of the people. The company has more than $10 million in signed contracts (often denoted using the term annual recurring revenue), according to two sources.
Decagon confirmed it has been in talks to raise additional funding, but said the round hasn’t been closed yet and the figures could change. Andreessen Horowitz and Accel did not immediately respond to comment requests.
Decagon is building customer support chatbots and agents— software that can autonomously perform basic tasks— for things like answering questions on how a product works, processing refunds and canceling subscriptions. Companies like Notion, Bilt, Duolingo, Substack and Rippling use Decagon customer support chatbots.
The latest investment comes less than a year after Decagon raised a $65 million Series B round led by Bain Capital Ventures at a $650 million valuation, Forbes reported in October. The new investment would bring Decagon’s total funding to $200 million. The company, which was featured on the Forbes AI 50 list, was cofounded by 27-year-old CEO Jesse Zhang and Ashwin Sreenivas in 2023 after the duo surveyed dozens of companies to find the problem that could best be solved with AI.
Customer support, a tedious but pivotal part of any business, has long been targeted for automation as a means to save millions in labor costs by eliminating largely outsourced customer support teams. Credit card provider Bilt downsized its customer support team from hundreds to 65 with Decagon, Forbes reported last year. And fitness giant ClassPass, which uses Decagon’s AI agents to carry out 2.5 million conversations with its customers, has reduced its customer support costs by 95%, according to The Information. “We think AI agents can be 10x employees,” Zhang told Forbes in March.
Under the hood, Decagon’s agents are built on the most advanced models from OpenAI, Anthropic and Cohere, and trained on internal data like how-to blogs, manuals and past customer service conversations. Staff rate and review responses generated by the AI to improve it. In February, Decagon partnered with audio generation startup ElevenLabs to create voice agents to have more natural, human-like conversations with customers.
Customer service software is a competitive market and Decagon is up against companies like $4.5 billion-valued Sierra, helmed by former Salesforce co-CEO and OpenAI board Chairman Bret Taylor, and Salesforce, which has its own AI agents for customer support. The products themselves are also increasingly identical to each other and businesses regularly do what are called bake offs to pit one AI tool against another to see how well they perform on the job.
“Decagon claims they win almost every time they go up against Sierra head to head,” said one investor involved in the talks. “Ultimately, they are both going to have exactly the same product and it will really be a race to see who can solve a higher percentage of the support calls with the best accuracy.”
MORE FROM FORBES
