Zendesk’s newly launched AI Outcome Platform and the company’s expanded use of an outcome pricing model all anticipate customer service increasingly becoming an agentic AI operation.
Zendesk is preparing for the not-too-distant future when agentic customer service applications respond to personal agents as often as they support real people.
This week, Zendesk launched the AI Outcome Platform, a new offering that connects internal customer support tools with external AI assistants. The company also announced its expanded use of an unlimited-seats, outcome-focused pricing plan that replaces traditional seat-based fees.
Shashi Upadhyay (pictured), Zendesk president of product, engineering and AI, predicted in an interview with CRN that in just a couple of years more than 50 percent of the traffic around customer service within businesses and organizations will be agent-to-agent.
“We’re seeing a pretty big shift in the world of customer experience and, specifically, in customer service,” Upadhyay said. “The big shift is that AI agents have been taking over, have become more and more capable, so they’re able to actually take over a lot of routine—and sometimes even more complex—tasks. The interactions shift from people-to-people to AI-to-AI.”
He noted that a significant driver is the growing use of personal agents such as Instinct and Meta Muse to handle personal tasks.
“Our view is that everyone in the world is going to have a personal agent who’s going to do real work for them,” Upadhyay said. “Real work means everything from, you know, booking flight tickets, to getting hotel reservations, to shopping for them—solving problems for them.” That will include contacting call centers and service desks to resolve billing errors, return merchandise, update personal information, and so on.
And that will include seeking assistance from systems like Zendesk. “So, it’ll be from personal agents being asked to do a job, and then a company’s agents, the service agents, actually doing the job,” he said.
The Outcome AI Platform, what Upadhyay described as an evolution of Zendesk’s flagship platform, “brings people, AI agents, apps, teams, everything together to better understand context—the intent of the human who wants a problem solved—complete work across the system, verify work across the system, and then learn from them,” the executive said.
The Outcome AI Platform incorporates a number of new and existing Zendesk technologies. Most prominent is the Open Service Directory (OSD) that enables personal AI assistants to locate brands’ verified support channels, regardless of whether or not the brand uses Zendesk. OSD also includes an open specification and public directory.
Also new in the platform is Zendesk Bridge, which connects external AI agents and interfaces, including ChatGPT, Claude, Microsoft Copilot and custom internal bots, directly to Zendesk’s service platform, including its AI, service context, workflows and actions. That, according to the company, allows customers to resolve issues using Zendesk data without leaving the AI tools they already use.
Also new are specialized AI agents for resolving complex customer and employee tasks using deep domain logic, and the Self-Improving Zendesk feature that helps teams continuously identify and resolve root-cause problems.
“Zendesk’s Outcome AI Platform is a natural extension of the work we’ve been doing together,” said Allison Johnson, Amazon Web Services director of Americas Technology Partners, in an email response to CRN questions. Zendesk and AWS maintain a deep development and go-to-market alliance under a Strategic Collaboration Agreement (SCA) announced in December 2025.
“Zendesk customers already use Amazon Bedrock and Amazon Connect to automate up to 80 percent of customer service interactions, and the Outcome AI Platform builds on that foundation as AI moves from answering questions to completing work,” Johnson said.
Outcome-Based Pricing Replacing Per-Seat Pricing
Zendesk’s Upadhyay said the traditional per-seat pricing model for software doesn’t work in the agentic world and the company previously began using outcome-based pricing for some of the company’s autonomous AI agent product offerings.
“The unit of measurement is going to be outcomes,” he said. “The platform itself is focused on outcome. First of all, it aligns us as providers much more closely with our customers and our partners. And it also helps customers achieve business results more directly because they can hold us responsible for helping them improve metrics like [customer] retention, conversion, growth, lifetime value—things that they care about.”
Johnson at AWS agrees: “AWS has long believed that customers should be able to connect what they spend on AI to the results they actually get. When AI does real work, the fairest way to measure its value is by what it accomplishes, not by how many people have a license.”
“That is the thinking behind our guidance on agentic software…that per-seat pricing was never designed to capture the value agents create, and that the right tools can tie an agent’s activity to real outcomes. Zendesk’s approach is a good example of pricing that follows that logic,” she told CRN.
Zendesk users “want one integrated experience,” Upadhyay said of the new platform. “Because the thing they care about at the end of the day—the customer doesn’t care about software, they don’t care about tech—what they care about is, are they getting the outcomes or not? And those outcomes are better achieved if it’s all integrated and works well together.”
Channel Implications
Upadhyay says all this is creating opportunities for channel partners.
“The thing we hear from all our customers is they are struggling to make a transition from a non-AI world, the one they had grown up in, to a world where AI is the center of their business operations. So they need help every step of the way,” he said.
Clients need expertise in understanding how AI will transform their businesses and operations, Upadhyay said. They need help developing and executing a transformation plan, defining the expected outcomes, and developing the means to measure those outcomes.
These needs create demand for new services “and those are opportunities for channel partners, SIs [systems integrators], other ecosystem partners—especially those who have built the trusted relationships with their customers over a period of time, to be those change agents to help them with this transition…The channel world is extremely good at doing that,” Upadhyay said.
“Zendesk’s Outcome AI Platform reflects the continued innovation and customer-focused approach that underpin our strategic relationship,” Johnson at AWS said. “By enabling organizations to scale AI-powered customer experiences, leverage shared customer context, and measure meaningful business outcomes, the Outcome AI Platform reinforces the opportunities our collaboration is designed to unlock.”
“We see significant potential to continue building on our relationship with Zendesk to help customers accelerate AI adoption, improve operational efficiency, and deliver better customer experiences,” she said.
