Pendo launches platform to analyze AI agent performance
The company said its new platform will allow businesses to track agentic AI software and optimize performance.
Pendo on Tuesday launched Agent Analytics, which the company touted as the first platform for measuring and optimizing agentic AI software performance.
The Agent Analytics platform enables businesses to measure AI agent performance as they would human employees and to track the adoption of agentic software as they would SaaS applications, the company said.
The platform will provide metrics and reports tracking in-house and third-party agent behavior in concert with traditional software.
Pendo CEO and co-founder Todd Olson said Agent Analytics was spurred by conversations with customers trying to get a handle on new agentic AI tools.
"What we heard loud and clear from the companies we work with is that there is an enterprise-wide mandate to start experimenting and trying out new agentic technologies," he said.
As they race to adopt AI tools, many enterprises are placing high hopes on agentic AI to provide return on investment, a hope that remains elusive. Almost half -- 48% -- of the 500 tech executives surveyed by EY in April are in the process of adopting or have fully deployed agentic AI stems. And 43% of respondents said more than half of their total AI budgets have been earmarked for agentic AI.
The quest for AI ROI
A 2025 C-suite Series IBM report surveying 2,000 CEOs found that only 25% of AI initiatives brought expected returns on investment in the last three years. Those same CEOs said they expected a positive ROI through efficiency and cost saving investments by 2027.
"Enterprises are investing heavily in agentic AI," said Pradeep Sanyal, AI and data leader at Capgemini. "Despite this surge, ROI remains elusive… many leaders admit they still lack the tools to track impact or even define success."
The difficulty lies in how agentic AI behaves, according to Sanyal. "These digital agents are autonomous, goal-seeking and often act like digital employees," he said "That makes their performance difficult to measure using legacy analytics tools. Enterprises need to know not just whether their agents are running, but whether they are delivering value."
Olson said Agent Analytics can give executives answers that lead to action.
"The intent is to help answer, 'What's the ROI?' And maybe you're not getting that initially," Olson said. "We envision companies using this product to iterate and update the agents until you reach a point where you are getting that ROI."
Arun Chandrasekaran, vice president and analyst at Gartner, said the release is a win for AI governance, ROI measurement and automation efficiency.
"The ability to track adoption and usage patterns, highlighting where automations succeed or need improvement, is important," he said in an email interview. "These solutions provide the governance and insights needed to continuously optimize and scale AI-driven workflows."
Sanyal said Pendo's release is a sign that agentic AI is maturing as a business use.
"It's not just a useful tool," he said. "It is a critical step toward scaling AI responsibly and capturing the returns enterprises have been chasing."
Shane Snider, a veteran journalist with more than 20 years of experience, covers IT infrastructure at Informa TechTarget.
