Autosana Inc., a startup building an agentic artificial intelligence platform for mobile and web app quality assurance, said today it raised $3.2 million in early-stage funding.

Front-end user interface and user experience has long been one of the most complex frontiers in testing on mobile and web.

In an interview with SiliconANGLE, Autosana co-founders Yuvan Sundrani (pictured, right) and Jason Steinberg (left) said the platform makes testing UI features less painful by turning codebases and natural-language prompts into test suites and automated flows across mobile and web.

Sundrani and Steinberg met in middle school and launched multiple mobile-first companies together, so they have seen the mobile development pain point together. Steinberg framed mobile as “the most unsolved problem” and the most direct. The web has numerous frameworks, whereas mobile hasn’t seen the same ecosystem attention.

“With tools like Cursor and Claude Code and all these coding agents, it’s making creating features so much faster, but 70% of our job was now QA,” explained Steinberg.

The advent of AI development tools has compressed development time, creating a situation where teams spend less time at the keyboard shipping new features. However, that speed is shifting the burden to verification. The weight is felt most sharply by small teams without dedicated QA, where regression testing — verifying the new features are bug free when they go live — becomes a serious bottleneck.

That pressure shows up at the front end, where modern apps absorb breaking changes constantly — and QA either breaks outright or misses issues that frustrate end users.

From spreadsheets to agents

The founders said the platform closes the loop between development cycles by acting like an engineer working long before code reaches production. Customers can hand the system a codebase and specifications, and Autosana can generate tests from there.

The system caters to a range of testers: engineers who want an agentic setup connected to their tooling and less technical manual testers who want a natural-language workflow in a dashboard they can “talk” to without living in the codebase. “We bring this power to both technical and non technical folks,” said Sundrani.

Manual testing in particular hits hard when it comes to regression testing. “You can kind of think of us like a glorified spreadsheet, rather than having someone manually do all of this, our agent actually just goes through every single case for you,” added Steinberg.

In the manual-testing world, spreadsheets become giant checklists of what needs to be tested, how it should be tested, and what happened, often whether it passed or failed or produced unexpected behavior. Unlike rules-based automation, agentic systems can “self-heal” around UI change because they can reason through visual shifts, even when a button moves, disappears, changes function, or a dropdown behaves slightly differently.

“Autosana is basically the agentic QA platform for iOS, Android and web apps… able to turn natural language into end-to-end testing,” said Sundrani.

The company said it currently powers QA for 100 million daily active users and claims revenue growth averaging more than 100% month-over-month since launching in June 2025, though it did not provide a baseline.

Autosana recently added web functionality to its agentic QA platform capabilities, and with this funding the company is not sitting on its laurels.

Sundrani and Steinberg said they’re betting that the familiar world of brittle end-to-end scripts is headed for retirement, replaced by AI agents that can see the interface, understand intent and keep up with constant UI change the way a human tester would — only at machine scale. In their view, Autosana is the start of a broader shift: a single, agent-driven QA platform that expands from mobile into web and desktop, then outward into richer validation.

Photo: Autosana