Zendesk is on track to generate roughly $400 million in AI revenue by year's end, a dramatic leap from virtually nothing two years prior, according to chief executive Tom Eggemeier. The bulk of this growth stems from AI agents that operate on a pay-per-resolution model, where customers are charged only when a case reaches closure.

Eggemeier, who has led Zendesk since late 2022, previously served as a partner at Permira, overseeing its Menlo Park operations. His career includes more than a decade at contact centre software firm Genesys, where he reached the rank of president. He holds a law degree from the University of Chicago and studied history at the University of Dayton.

Speaking at HumanX in Amsterdam on 23 September, Eggemeier appeared on the TNW Podcast, the third episode following conversations with Yahoo's Jim Lanzone and Zoom's XD Huang.

$3,105 with Muse

Meta's Muse personal assistant generated tangible savings for Eggemeier's household, following Meta's challenge to users to save up to $1,000. According to Eggemeier, "I saved over the last five days $3,105." The tool reduced his family's internet and TV expenses by approximately $2,000 annually and trimmed mobile costs by around $1,000 per year. Muse handled roughly 95 percent of the negotiation work, with Eggemeier making just one phone call himself. He is not yet part of the early testing phase where Muse places calls independently.

Eggemeier noted that European consumers may not experience comparable savings, given that internet and mobile pricing in the US significantly exceeds European rates. He has been experimenting with both Muse and Instinct, another personal assistant tool, in recent weeks.

In his daily work, Eggemeier relies on ChatGPT for multiple hours each day, using it for research, customer engagement and speech preparation. Working alongside his team, he developed code using Claude Code that is now running in production systems at Zendesk, after the team reviewed it three times. He has also created a basketball application with his son.

Paying for outcomes

Zendesk introduced pricing based on automated resolutions in August 2024. The company has since piloted a revised approach with approximately 100 customers that eliminates the traditional seat-based model. Under this new structure, Zendesk charges per resolution, whether handled by human agents or AI systems. The company plans to make this option available shortly to interested customers, Eggemeier stated. He also presented the case against seat-based pricing during a HumanX presentation on the same day.

AI now resolves or supports more than 50 percent of Zendesk tickets, according to Eggemeier, a figure that encompasses self-service solutions, AI agents and copilots combined. AI agents independently manage a smaller portion of this total. His most successful customers achieve automation rates between 80 and 90 percent, with satisfaction levels exceeding those of human-handled interactions.

A utility company operating in Brazil deployed Zendesk's collections agent to generate approximately $5 million in additional revenue within months. More than 80,000 companies currently use Zendesk, and Eggemeier estimates that between one and two billion people interact with the platform annually without realizing it.

A human on call

Eggemeier anticipates that conversations between personal AI agents and company-operated agents will constitute more than half of all interactions within two years. He described this prospect as somewhat unsettling, while emphasizing that humans must always remain accessible. Zendesk advises its customers to maintain this option, he noted.

When asked about companies using AI to reduce their workforce, Eggemeier argued that both businesses and governments should invest in training people to work effectively with these technologies. Describing himself as optimistic, he drew parallels to job transformations following the industrial revolution and since the 1980s.

I can't see beyond five or 10 years. There's going to be absolutely a need for humans.

Tom Eggemeier

For young people, including his own two children attending university, Eggemeier recommends developing strong communication, analytical and interpersonal abilities alongside AI literacy.

Source: The Next Web