One of the hottest areas in the generative AI space is customer support, which isn't surprising given the technology's potential to increase scale while reducing contact center costs. Critics argue that generative AI-powered customer support technology could lower wages, lead to layoffs, and ultimately result in an error-prone end-user experience. Meanwhile, proponents argue that generative AI augments workers rather than replaces them, freeing them up to focus on more meaningful tasks.
Jesse Chan is an advocate — though a little biased, of course. Together with Ashwin Sreenivas, Chan co-founded Decagon, a generative AI platform that automates various aspects of customer support channels.
Chan understands how competitive the AI-powered customer support market is, with competitors including tech giants like Google and Amazon, as well as startups like Parlore, LittelAI, and Cognigy (which recently raised $100 million). By one estimate, the sector could be worth $2.89 billion by 2032, up from $308.4 million in 2022.
But Chan believes both Decagon's engineering expertise and go-to-market approach give the company an edge. “When we first started, the general advice was not to pursue customer support because it was too competitive,” Chan told TechCrunch. “Ultimately, what worked for us was to aggressively prioritize what customers wanted and stay focused on what they could get value from. That's the difference between a real business and a flashy AI demo.”
Both Chan and Sreenivas have backgrounds in technology and have worked at both startups and large tech companies: Chan was a software engineer at Google, then a trader at market-making firm Citadel, before founding Lokey, a social gaming platform that was acquired by Pokémon Go developer Niantic in 2021; Sreenivas was a deployment strategist at Palantir before co-founding computer vision startup Helia, which he sold to unicorn Scale AI in 2020.
Decagon sells primarily to large enterprises and “fast-growth” startups, developing customer support chatbots. Powered by first- and third-party AI models, the bots can be fine-tuned and can pull from a company's knowledge base and historical conversations with customers to better understand the context of an issue.
“When we started building, we realized that a 'human-like bot' has a lot to offer, as human agents are capable of complex reasoning, behavior, and post-conversation analysis,” Zhang said. “When we talk to customers, it's clear that while everyone wants to improve operational efficiency, it shouldn't come at the expense of customer experience. Nobody likes chatbots.”
Decagon uses generative AI technology to answer customer questions and more. Image Credit: DecagonImage Credit: Decagon
So how is Decagon's bot different from a traditional chatbot? According to Zhang, the bot learns from past conversations and feedback. And perhaps more importantly, it can integrate with other apps to take actions on behalf of customers and agents, like processing refunds, categorizing incoming messages, or helping create support articles.
On the backend, companies will be able to analyze and control Decagon's bots and their conversations.
“Human agents can analyze conversations to spot trends and identify areas for improvement,” Zhang says. “Our AI-powered analytics dashboard automatically reviews and tags customer conversations, identifying themes, flagging anomalies, and suggesting additions to the knowledge base to better address customer inquiries.”
Now, generative AI has a reputation for being less than perfect and, in some cases, ethically questionable. What would Chang say to companies that worry that Decagon's bots might tell someone to eat glue or write articles full of plagiarized content, or that Decagon might use their data to train their models?
Essentially, he says, there's no need to worry. “It's important that we provide our customers with the guardrails and oversight they need for their AI agents,” he said. “Ensuring consistency across many conversations is no easy task. We achieve this by fine-tuning our models and providing our customers with a rigorous testing framework to ensure their AI agents are behaving as expected.”
While Decagon's technology faces the same limitations as other generative AI-powered apps, it has recently attracted big-name clients like Eventbrite, Bilt, and Substack, helping the company reach breakeven. High-profile investors including Box CEO Aaron Levie, Airtable CEO Howie Liu, and Lattice CEO Jack Altman have also joined the venture.
To date, Decagon has raised $35 million in seed and Series A funding, with participation from Andreessen Horowitz, Accel (which led the Series A), A*, and entrepreneur Elad Gil. Chan said the funding will be used for product development and expanding Decagon's San Francisco-based workforce.
“The main challenge is that customers equate AI agents with previous-generation chatbots that don't actually get the job done,” says Zhang. “The customer support market is saturated with old chatbots that have lost consumer trust. This new generation of solutions needs to cut through the noise of existing products.”