Zendesk Charges AI Customers Only for Resolved Issues
Pay-per-result pricing for customer-service AI is announced, not yet broadly available, and follows a similar OpenAI trial with select customers.
Updated on Nov. 11, 2026 in Artificial Intelligence
Live Poll
Do you trust companies' AI customer service to actually solve your problems?
Zendesk announced a pay-per-result pricing model for its AI through a new Outcome AI Platform at its AI Summit in San Francisco. Customers set an Outcome Budget and pay for each eligible issue the AI resolves or task it completes, and Zendesk checks each result for quality and completeness. The pricing model is announced rather than a settled, broadly available product.
Why it matters
The model ties AI cost to verified outcomes rather than to seats, a shift Zendesk argued for earlier this year alongside Sierra. OpenAI has started trying a similar approach with some customers, suggesting pay-per-result pricing is becoming a competitive pattern in customer-service AI.
The pricing is tied to verified outcomes: Zendesk checks each AI result for quality and completeness before a customer is charged for it. The source provides no benchmark figures for resolution accuracy or cost per resolved issue.
The players
Zendesk
A customer-service software company whose Outcome AI Platform, Bridge integration layer and Open Service Directory connect AI assistants to businesses' support systems.
OpenAI
The developer of ChatGPT, which has started testing a similar outcome-based pricing approach with some customers.
Sierra
An AI customer-service company that argued alongside Zendesk earlier this year that AI agents should be paid per result.
IDC
A technology market research firm whose research director, Tapan Patel, commented on the benefits of linking cost to verified results.
The details
Zendesk Bridge lets businesses connect AI agents and apps, such as ChatGPT, to Zendesk's systems, and every action passes through Bridge. Bridge checks permissions and company rules and asks for human approval where needed. Zendesk's Open Service Directory is a public list where brands register official support channels for AI assistants, and it does not access private customer data or take any action itself.
Timeline
Earlier this year, Zendesk and Sierra argued that AI agents should be paid per result, not per seat.
On Tuesday, November 10, 2026, Zendesk announced the Outcome AI Platform at its AI Summit in San Francisco.
The Open Service Directory is launching soon, with no specific date given.
The Tech Race
OpenAI has started trying a similar outcome-based approach with some customers, placing Zendesk's model in direct competition with the largest AI developer. The move positions Zendesk against both AI-native customer-service startups like Sierra and the model providers now selling into the same support workflows.
Businesses on Zendesk can set an Outcome Budget and pay only for eligible issues the AI resolves, but availability for general customers is not specified beyond early access for Bridge. Pricing for customers outside early access has not been detailed in the source.
The takeaway
Pay-per-result pricing moves the cost question from seats to verified outcomes, a trajectory worth tracking as OpenAI and others test similar models. Watch for the Open Service Directory launch and the first named customer results from Bridge in early access.
What happens next
The Open Service Directory is launching soon, and Zendesk Bridge remains in early access, so the first public availability of both is a near-term milestone to watch.
Further reading
For broader context on how AI agents are reshaping business software, explore our Artificial Intelligence coverage.
Live Poll
Do you trust companies' AI customer service to actually solve your problems?
