Testlio has launched AI Agent Testing services aimed at companies deploying AI agents in customer-facing and transaction-heavy environments.
The offering combines human-in-the-loop testing with AI-based analysis to assess how AI agents behave in real-world workflows. It is intended to test agentic systems before they reach customers, with a focus on failures that could create compliance, financial, or reputational risk.
Testlio is entering a market where businesses are adopting AI agents more widely but still face reliability concerns. Citing research, the company said 79% of enterprises have implemented AI agents at some level, while most AI agents complete tasks successfully on the first attempt only 24% of the time.
Risk focus
The service targets four areas that companies often struggle to assess through automated testing alone: how agents respond to unpredictable real-world cases; whether subtle behavioural issues are missed by automation-only reviews; whether data passes accurately between workflow steps; and how quickly teams can spot issues when software is released at speed.
According to Testlio, results from these assessments feed into LeoPulse, its internal AI confidence score. The score is used to support release decisions and track performance over time.
Testlio is pitching the service particularly to sectors where errors can carry high costs. Finance, eCommerce, media, and software are among the industries where it already provides quality assurance work and where a malfunction can trigger customer complaints, regulatory scrutiny, or lost revenue.
Summer Weisberg, Chief Executive Officer of Testlio, said the new service builds on the company's existing work in payments testing and applies that experience to AI-driven workflows.
"Testlio is uniquely positioned for this moment, combining years of proven global payments testing with this new AI agent validation service," Weisberg said.
"As enterprise AI agents evolve from handling simple customer queries to executing complex agentic payment workflows, where AI is granted digital wallets to independently authorize transactions, organizations need a partner that goes beyond automated simulation. They need access to humans with real, localized payment methods to act as a safety net, alongside experts specializing in AI agent evaluation. That's exactly what Testlio provides," she said.
Payments backdrop
The emphasis on payments reflects one of the more sensitive uses of agentic AI, where software is expected not just to recommend actions but to initiate them. In such cases, testing must account for different devices, local payment methods, and user behaviour that may not follow fixed scripts.
Testlio said it uses a global network of AI-certified testers to run what-if scenarios in local markets. It argues this type of testing can reveal failures that would not appear in controlled environments or standard automation frameworks.
Dean Hickman-Smith, Chief Revenue Officer of Testlio, said businesses will need a different approach as autonomous systems become more common.
"Autonomous agents require an entirely new testing playbook," Hickman-Smith said.
"The future of AI testing is not automation alone. It's AI combined with localized human judgment, real-world validation, and continuous feedback loops that enterprises can trust at scale," he said.
Testlio said its wider client roster includes PayPal, NBCUniversal, Strava, Adyen, Dlocal, Uber, and Solidgate. Those relationships are central to its argument that it already has experience in high-risk digital workflows and access to the local payment rails AI agents may eventually use.
One recent example cited by the company involved Solidgate, a payment orchestration platform operating in more than 150 markets. Testlio said the work helped the client test embedded payment systems against regional requirements and achieve a production issue resolution rate of more than 95%.
Stefan Dzhenkov, Head of Alternative and Local Payments at Solidgate, described the value of in-market testing for payment flows.
"Working with Testlio allowed us to catch critical issues we wouldn't have seen otherwise. The feedback from in-market testers gave us confidence that our payment flows work as expected, and we could address problems before they reached our customers," Dzhenkov said.
Testlio operates as a fully remote company with team members in more than 150 countries. It positions its global tester network as a key part of its approach to quality assurance for software releases and AI-driven workflows.