Presented by Zendesk
Shashi Upadhyay, president of engineering, AI, and product at Zendesk, says agentic AI is currently transforming three key areas of work — creative, coding and support. But he notes that support presents a different challenge.
"Support is special because you are putting an autonomous AI agent in front of your user," Upadhyaya says. "You have to trust that it’s going to do the right thing for the customer and the customer. Every step forward in AI should make the service more reliable for both users and human agents."
Zendesk, recently named a leader in this 2025 Gartner Magic Quadrant Started implementing AI agents a year and a half ago, for a CRM customer engagement center. Since then, they have seen that AI agents can solve about 80% of all incoming customer requests on their own. For the remaining 20%, the AI agent can hand it over to a human to help solve more complex problems.
"Autonomous AI agents work 24/7, with no waiting or queuing time. You have a problem. He immediately. They provide the answer. All this adds up," He says. "Not only do you get higher resolutions, higher automation, but you can improve CSAT at the same time. Because 80% is such a smart number, and the results are so solid, we think it’s only a matter of time before everyone embraces this technology. We’re already seeing it across the board."
General AI-powered insights with constant testing, integration of advanced models like ChatGPT5, and a major upgrade to its analytics capabilities and real-time, Hyperarch, an AI-native analytics platform for the company’s efforts to advance its quality of use, depth of insight, and value for organizations of all sizes.
Designing, testing, and deploying improved agents
"Especially in a support context, these important AI agents deal consistently with the company’s brand, policies, and regulatory requirements that you may encounter." Upadhyaya says. "We consistently test every agent, every model across all our customers. We do this before we release it and we do it after we release it, in five categories."
Those categories — automation rate, execution, accuracy, latency and safety — form the basis of Zendesk’s ongoing benchmarking program. Each model is scored on how accurately it solves the problems, how well it follows instructions, how quickly it responds, and whether it stays within clearly defined guards. The goal isn’t just to make AI faster — it’s to make it reliable, responsive, and aligned with the standards that define great customer service.
This testing is powered by Zendesk’s QA Agent – an automated monitor that keeps a constant eye on every conversation. If an exchange starts to veer off course, whether in tone or accuracy, the system immediately flags it and alerts a human agent to step in. It’s an extra layer of assurance that keeps the customer experience on track, even when AI is running the first line of support.
GPT 5 for next-level agents
In the world of support and service, this move from basic chatbots that answer basic questions or solve uncomplicated problems to agents who actually take action is groundbreaking. An agent that can understand that a customer wants to return an item, verify if it’s eligible for a return, process the return, and issue a refund is a powerful upgrade. With the introduction of ChatGPT5, Zendesk recognized an opportunity to integrate this capability into its resolution platform.
"We worked Very closely with Openei Because the GPT-5 model’s capabilities were a huge improvement, from being able to answer questions, to being able to reason and take action," Upadhyaya says. "First, it does a much better job of solving problems autonomously. Secondly, it’s much better at understanding your intent, which improves the customer experience because you’re understood. Last but not least, it has 95% plus reliability when executed correctly."
Those benefits slide across Zendesk’s AI agents, Copilot, and App Builder. GPT5 reduced workflow failures by 30 percent, thanks to the ability to adapt to unexpected complexity without losing context, and reduced fallback increases by more than 20 percent, with more complete and accurate responses. The result: faster resolutions, less hands-off, and an AI that behaves more like a seasoned support professional than a scripted assistant.
In addition, GPT5 is better at handling ambiguity, and is able to clarify ambiguous customer input, which improves routing and increases automated workflows in more than 65 percent of conversations. It has greater accuracy in five languages, and makes agents more productive with more concise, contextual responses that align with tone guidelines.
And in App Builder, the GPT-5 delivered 25% to 30 percent faster overall performance, with more immediate iterations per minute, leading to faster App Builder development workflows.
Bridging the Analytics Gap
Traditionally, support analytics has focused on structured data—the kind that fits neatly on a table: when a ticket was opened, who handled it, how long it took to resolve it, and when it was closed. But the most valuable insights often reside in unstructured data — itself spread across conversations, email, chat, voice and messaging apps like WhatsApp.
"Consumers often don’t realize how much intelligence goes into their conversations." Upadhyaya says. "What we’re emphasizing with analytics is how we can improve the entire company with the insights that sit in the supporting data."
To surface these deep insights, Zendesk turned to Hyperarc, an AI-native analytics company known for its proprietary Hypergraph engine and Generative-A-powered insights. The acquisition gave new life to Discover New Life, Zendesk’s analytics platform, transforming it into an innovative solution capable of integrating structured and unstructured data, supporting conversational interfaces, and drawing on persistent memory to use past conversations as context for new questions.
"Your support conversations are telling you everything that isn’t working in your business today, all that information sitting in the millions of tickets you’ve collected over time." Upadhyaya says. "We wanted to make it fully visible. Now we have this genius AI agent that can analyze all of this and come back with clear recommendations. It doesn’t just improve support. It makes the whole company better."
This visibility now translates into actionable intelligence. The system can identify where problems are most persistent, identify patterns behind them, and suggest ways to fix them. It can even predict problems before they happen. During high-pressure events like Black Friday, for example, it can analyze historical data to flag recurring problems, predict where new disruptions might appear, and recommend preventive measures—turning reactive support into a proactive strategy.
"That’s where the hyperarche shines," Upadhyaya says. It doesn’t help you understand the past – it helps you plan better for the future."
By integrating Hyperarch’s AI-local intelligence, Zendesk is moving customer service toward continuous learning—where every interaction builds trust and accelerates efficiency, setting the stage for AI that can see what’s coming next.
Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and is always clearly marked. For more information, contact sales@ventorbet.com.