Applause doubles down on AI testing as software quality gaps widen

Michiel Willems
April 17, 2026

US-based software testing firm Applause is accelerating its shift into AI-driven quality assurance, appointing Aatish Salvi as chief technology officer as many bankss and firms struggle to validate increasingly complex, AI-powered systems.

The move comes as banks and financial institutions face mounting pressure to test applications built on generative and agentic AI, where traditional QA methods are proving insufficient.

Applause, known for its managed testing services and global crowdtesting model, is positioning itself at the centre of this shift toward AI assurance and real-world validation.

“Generative and agentic AI are the latest shift, and one of the most complex,” explained CEO Chris Malone. “While these technologies are changing how software is built, they also introduce new risks that traditional testing approaches can’t fully address.”

Salvi will lead Applause’s global technology and product organisation, overseeing a roadmap focused on expanding AI-enabled testing capabilities and automation.

His appointment reflects a broader transition across the software testing industry, where QA is evolving alongside faster release cycles and increasingly autonomous software behaviour.

“With AI-driven software development accelerating, organisations need to stay ahead of an increasingly complex testing landscape,” Malone stressed. “As we expand our capabilities and coverage, having the right technical leadership is critical.”

Hybrid testing models

At the core of Applause’s strategy is a hybrid approach that combines automation, artificial intelligence and human testers operating across real-world environments.

The company argues this model is essential as AI systems introduce unpredictable and non-deterministic behaviours that are difficult to capture through conventional testing.

“Leading brands come to us because we’ve always helped them test what’s next, from early mobile and digital platforms to today’s AI-driven applications,” Malone said.

“Our model validates how applications perform in real-world conditions with the speed and scale of AI, so teams can move faster with confidence and deliver high-quality experiences that retain customers and drive growth.,” he added.

“Teams are pushing AI into production before they’ve figured out how to properly test it.”
– Chris Sheehan

Salvi echoed the growing stakes for enterprises deploying AI into production environments, particularly in sectors such as financial services where failures can have immediate commercial and regulatory consequences.

“If product releases fail, customers and revenue are on the line,” he noted. “We’re experiencing unprecedented technological transformation, and for our clients, the stakes are incredibly high.”

As part of this push, Applause is expanding services aimed at testing AI systems themselves, including model evaluation, domain expert validation and red-teaming exercises designed to identify risks, inaccuracies and unsafe outputs before deployment.

Quality falls behind

The company’s latest State of Digital Quality in Testing AI report underscores the scale of the challenge facing QA teams. While AI adoption is surging across enterprises, quality assurance is struggling to keep pace.

According to the report, just over half of organisations have already released AI-powered applications and features, yet more than half of AI initiatives fail to reach full production due to integration challenges, cost constraints and quality risks.

“AI development isn’t slowing down, and quality is falling behind,” said Chris Sheehan, EVP of High Tech and AI at Applause.

“Teams are pushing AI into production before they’ve figured out how to properly test it. That’s why we’re seeing more failures and more risk reaching users.”

The data points to rising user-facing issues, including hallucinations, misunderstood prompts and unreliable outputs, as well as a growing disconnect between rapid deployment cycles and the ability to validate systems at scale.

“AI adds speed and scale, but human evaluation is what earns trust, you need both,” Sheehan said. “The companies getting it right combine AI and domain expertise to evaluate and fine-tune their systems, ensuring outputs are more relevant, accurate and inclusive.”

Despite advances in automated testing techniques, human evaluation continues to play a central role in AI validation.

Applause found that 61% of organisations rely on human input to assess AI performance, even as newer approaches such as LLM-as-judge gain traction.

“Testing AI isn’t just about accuracy, it’s about evaluating complex, multimodal outputs at scale,” said Chris Munroe, VP of AI Programs at Applause.

“LLM-as-judge systems are becoming an important part of that process, but they can’t operate in isolation. Without human oversight, you risk reinforcing the same blind spots you’re trying to detect.”

Munroe added that structured red teaming and rigorous evaluation frameworks are essential to ensure systems are safe and reliable before deployment.

“In addition to human-led evals and fine-tuning, structured red teaming by both domain experts and generalists is essential,” he continued. “So is ensuring evaluation rigor, without it, organizations risk scaling systems they don’t fully understand or control.”

For banks and financial services firms, the implications are significant. As AI systems become more embedded in core operations, from customer service to decision-making, software testing is shifting from a technical function to a strategic control point tied to resilience, trust and regulatory readiness.

Applause’s latest moves highlight how testing providers are expanding beyond traditional QA into broader AI governance and assurance roles, helping enterprises validate not just whether systems work, but whether they behave safely and predictably in real-world conditions.