Even minor glitches can damage a company's reputation and bottom line.
In an era when digital experience has become the front line of business reputation, Keysight Technologies has deepened the intelligence of its Eggplant testing platform by teaching machines to see — and judge — software the way a human user would, only faster. The February 2021 upgrade brings real-time computer vision and self-healing AI to the testing of high-speed iOS applications, addressing a quiet but consequential gap: the inability of traditional methods to catch failures at the speed modern software demands. From gaming to autonomous vehicles to financial trading, the stakes of a missed glitch are no longer merely technical — they are existential for the brands that ship the code.
- Modern users have zero tolerance for lag or failure, and applications in gaming, autonomous vehicles, and financial trading operate at speeds that expose every weakness in traditional testing approaches.
- Keysight's Eggplant DAI becomes the first AI-driven platform to integrate computer vision specifically for high-speed iOS testing, detecting visual changes on screen with millisecond precision.
- Self-healing test assets quietly absorb application changes on their own, sparing engineering teams the costly, repetitive work of manually updating test suites every time software evolves.
- Real user behavior — bounces, conversions, session duration, revenue impact — is now surfaced directly inside the platform, closing the loop between performance data and business outcomes.
- Expansion to additional operating systems is already planned, signaling that this is not a one-platform fix but the opening move in a broader strategy to own the intelligent testing lifecycle.
Keysight Technologies upgraded its Eggplant Digital Automation Intelligence platform in February 2021, introducing real-time computer vision designed to automate the testing of high-speed iOS applications with millisecond-level accuracy. The move targets a genuine vulnerability in modern software development: as user expectations for flawless digital experiences have risen, the applications most sensitive to failure — gaming engines, autonomous vehicle systems, financial trading platforms, medical devices — have outpaced the ability of conventional testing tools to keep up.
The platform now uses artificial intelligence to recognize and respond to visual changes on screen in real time, effectively replicating how a human tester perceives software behavior, but at machine speed and with machine consistency. Alongside this, Keysight introduced enhanced self-healing test assets — AI that monitors application changes and automatically updates image variants without manual intervention, reducing the ongoing cost of maintaining test coverage as software evolves.
The upgrade also brings richer analytics into the picture. Teams can now access data drawn from real user sessions — including bounce rates, conversions, load times, session duration, and revenue impact — through an improved reporting interface with better visualization and filtering. The intent is to connect testing outcomes directly to business performance, not just technical benchmarks.
Gareth Smith, general manager of Eggplant, described the enhancements as a response to a world where digital experience now determines competitive standing. Keysight, which posted $4.2 billion in revenue for fiscal year 2020 and serves industries from aerospace to semiconductors, plans to extend the computer vision capability to additional operating systems in the months ahead — positioning the platform as a full-lifecycle testing solution for enterprises that can no longer afford to ship software blind.
Keysight Technologies announced an upgrade to its Eggplant Digital Automation Intelligence platform that brings real-time computer vision into the testing process for high-speed applications. The enhancement allows organizations to automate the evaluation of iOS apps with millisecond-level precision, simulating how users will actually experience the software in the wild.
The problem the upgrade addresses is straightforward: in a world where users expect flawless performance from every app they touch, even minor glitches or delays can damage a company's reputation and bottom line. Applications that demand instantaneous response—gaming software, autonomous vehicle systems, financial trading platforms, medical devices—have no margin for error. Traditional testing methods struggle to catch performance issues at the speed these systems operate. Keysight's solution uses artificial intelligence and computer vision to recognize visual changes on screen and respond to them in real time, effectively teaching machines to test the way humans do, but faster and more consistently.
Eggplant DAI is the first AI-driven automation platform to integrate computer vision specifically for testing high-speed iOS applications. The company plans to extend this capability to other operating systems in the coming months. What makes this meaningful is not just the speed of detection, but the reduction in ongoing maintenance burden. The platform now includes what Keysight calls enhanced self-healing test assets—AI that automatically learns and adapts as the underlying application changes, adding new image variants without manual intervention. This cuts the cost of keeping test suites current as software evolves.
The platform also now surfaces interactive insights drawn directly from real user behavior. Organizations can access data on bounces, conversions, user demographics, session duration, load times, page impressions, revenue impact, and session analytics. The new reporting interface includes better visualization and filtering tools, giving teams the ability to connect performance metrics to actual business outcomes.
Gareth Smith, general manager of Eggplant, framed the upgrade as a response to rising user expectations. "Delivering digital experiences that delight now determines business success," he said, noting that the most demanding users will not accept even minor performance issues. The enhancements, he argued, demonstrate the platform's ability to automate the entire testing lifecycle with intelligence built in.
Keysight Technologies, which reported $4.2 billion in revenue for fiscal year 2020, serves enterprises, service providers, and government agencies across communications, aerospace and defense, automotive, energy, semiconductor, and general electronics sectors. The company's broader mission is to help organizations accelerate innovation and bring products to market faster and more cost-effectively. This Eggplant upgrade fits that mandate: by automating and intelligently managing the testing phase, organizations can catch problems earlier and ship with greater confidence.
Notable Quotes
Organizations now have a way to meet the needs of sophisticated users who will not tolerate even the slightest glitch or performance issue.— Gareth Smith, Eggplant General Manager
The Hearth Conversation Another angle on the story
Why does millisecond accuracy matter so much in app testing? Isn't that just engineering perfectionism?
It's not perfectionism—it's the difference between an app that feels responsive and one that feels broken. In gaming or trading platforms, a 100-millisecond delay can mean a missed trade or a lost game. Users don't think about milliseconds; they just feel whether something works or doesn't.
So the computer vision piece—that's the real innovation here. What does it actually do that traditional testing can't?
Traditional testing often works by checking specific code paths or comparing outputs to expected values. Computer vision watches what the user sees on screen and detects when something changes unexpectedly. It's testing the actual visual experience, not just the backend logic.
And the self-healing part—that sounds like it could be a double-edged sword. What if the AI learns the wrong thing?
That's a fair concern. But the idea is that it's learning visual variants of the same feature as the app gets redesigned or updated. It's not making judgment calls about whether something is correct—it's just recognizing that the same button might look slightly different in a new version and not flagging that as a failure.
Who actually benefits most from this? Is it the big tech companies or smaller teams?
Both, but differently. A large organization testing a complex app across many devices and user scenarios saves enormous time and money. A smaller team gets access to testing sophistication they couldn't afford to build themselves. The real win is that testing becomes less of a bottleneck.
What happens when this rolls out to Android and other platforms?
That's when it becomes genuinely transformative. Right now it's iOS-only, which is a meaningful limitation. Once it covers the major platforms, organizations can test their entire digital product line with the same intelligent, automated approach.