Co-founder Ashwin Sreenivas’s Decagon is a Leader in Customer Service Automation

Decagon, backed by successful venture investors, is among the startups rapidly automating white collar jobs

(Photo: Jesse Zhang, left, and Ashwin Sreenivas, co-founders Decagon.)

February 21, 2025

This publication, Global Indian Times (GIT) is published on Substack, a digital media platform for video, writing, podcasts, and other content.

Recently, GIT sought to use a domain name for publishing on Substack. There is no customer service telephone number to call. Instead, there is an Ask A Question button on the Substack dashboard. Previous attempts to solve service issues, by using such chatbot help on sites, were unsatisfactory. But, since there was no choice, we used the button on Substack and typed in: how to use a domain name on Substack? This time, in less than a minute, there was an answer with clear, step-by-step instructions, which we followed to link a domain name for GIT on the platform.

There are several thousand publishers on Substack who together have more than 35 million subscribers. For its customer service, the platform uses an artificial intelligence (AI) agent from Decagon. The AI agent, based on large language models, replaced a chat-bot built in-house by Substack. Decagon and Substack are both based in San Francisco, California.

“The pain of a bad customer service experience is universal. No one likes being placed on hold for hours, being transferred to a string of different agents, or emailing back-and-forth for days only to come to an unsatisfactory resolution,” notes a blog post on the site of Andreessen Horowitz, a Menlo Park, California, based venture capital firm which manages $45 billion, and early investor in Decagon.

Use of the Decagon AI agent has “shortened our time-to-resolution rate…and significantly enhanced our approach to support by automating routine tasks,” a Substack manager notes on a Decagon blog. As a result, the platform is “capable of handling and resolving more than 90% of user inquiries” and provide “an even higher level of service to publishers and their subscribers.”

Decagon’s AI agents manage customer interactions from start to end, including fetching data, taking actions, and reviewing conversations.

In addition to Substack, Decagon’s customers include ClassPass, a fitness membership which offers users access to gyms in more than 2,500 cities worldwide, and Fourthwall, which helps content creators on YouTube, Twitch and other platforms consolidate their various income streams.

Decagon was co-founded in 2023 by Jesse Zhang, also chief executive, and Ashwin Sreenivas, chief technology officer. Zhang, 30, sold his previous startup Lowkey to Niantic.

In 2019, Sreenivas, 30, founded Helia, which applied AI to real-time video, which he sold a year later to Scale AI. Earlier for a year, he worked as a strategist at Palantir Technologies New York office. Sreenivas earned a Masters, 2019, and a Bachelors, 2017, both in computer science from Stanford University. Zhang earned a degree in computer science from Harvard.

In October, Decagon raised $65 million in a funding round, including from Elad Gil, a venture investor who has backed more than 40 startups which grew into $1 billion plus valuations.

There are more than 47 startups, in the U.S., Western Europe, and India, who sell products which automate customer service work. The startups are armed with billions of dollars of funding from venture capital firms. While Decagon has several clients and is funded by successful venture investors, it may be a few years before it, or one of its rivals, emerge among the few winners with a large market share in AI use for automating customer service and other white-collar jobs.

Meanwhile, in the U.S. and some other Western countries, self-service AI chat-bots increasingly enable customers, of even startups like Substack, to follow instructions and solve technical, bill payment, and other service issues without the help of human service representatives.

Similarly, staff at major banks and other companies, in the U.S. and elsewhere, use AI tools to accurately and efficiently answer phone and online questions from customers. Since the use of AI tools enable the staff to tackle customer questions far more quickly, each of them now handles a far larger volume of calls than they did in 2022. So, the companies need fewer customer service staff.

The expanding use of Decagon and other AI tools is reducing demand for customer service staff in India, Philippines, and elsewhere. So far, there is no data on such job losses published by government agencies or trade groups in India.

There are around 4.5 million employed by information technology (IT) companies in India, including about 1.4 million in customer service and call center work, mainly for foreign companies, according to NASSCOM, the trade association of the Indian IT industry. This data though is four years old.

Many of those employed by IT companies in India are at risk of losing their jobs due to rising use of AI and other automation tools.

“AI is often seen as destroying jobs, but at Decagon, we believe the opposite. Our AI agents are enhancing jobs, not replacing them,” Jesse Zhang, CEO of Decagon stated in a company blog post. Yet the same Decagon blog post also notes that the company’s “customers have seen significant efficiency improvements, enabling them to serve customers better with fewer human resources…”

A post on the venture firm Andreessen’s site also apparently hints at cuts in labor costs while discussing Decagon’s business prospects: “there’s a constant tension between providing first-class, on-demand support while also remaining cost-efficient.”

For unique stories each week, SUBSCRIBE to Global Indian Times by sending your email address to: gitimescontact@gmail.com