Opkey.

Every enterprise running a modern cloud ERP knows the same quiet dread. A vendor pushes an update - Oracle, Workday, SAP, Salesforce - and somewhere in a finance or HR workflow, something breaks. Someone spends a weekend hunting for it. Multiply that by four update cycles a year across dozens of modules, and you get what Opkey's co-founder and CEO Pankaj Goel bluntly calls "testing hell."

Opkey is the company built to end it. Founded in 2015 by three longtime friends and former Adobe and Oracle engineers - Pankaj Goel, Avinash Tiwari and Lalit Jain - Opkey sells an AI-powered, no-code platform that tests, configures and maintains packaged cloud applications. Instead of QA teams hand-writing scripts that shatter every time a screen changes, Opkey discovers what needs testing, generates the tests, and - when the application shifts - heals the scripts automatically.

The pitch is deceptively unglamorous, and that is exactly the point. While much of the AI industry chases models that write essays and images, Opkey pointed its technology at the tedious, high-stakes plumbing of enterprise software. That focus has earned it more than 250 customers, roughly 70% of them in the Fortune 1000, and a $47 million Series B in August 2024 led by PeakSpan Capital.

What Opkey actually does

At its core, Opkey is a Cloud Application Lifecycle Management platform. It ships with a library of more than 30,000 pre-built tests spanning Oracle Cloud, Workday, SAP, Salesforce, Coupa and other enterprise systems, so many customers start automating on day one rather than building from scratch. Process mining discovers the business flows that actually run in an organization; impact analysis maps which of those flows an incoming update will touch; and self-healing scripts keep the whole suite alive as applications evolve.

The result is a shift in where testing risk lives. For most enterprises the terrifying moment is the ERP go-live. Opkey's entire design argues the opposite: the go-live is one day, but the hundreds of updates afterward are the marathon - and that is the phase it optimizes.

Argus, and the case for a small, expert model

In 2024 Opkey introduced Argus, an enterprise-specific AI model of roughly 8 billion parameters, trained on terabytes of proprietary functional and technical ERP knowledge. It is deliberately small for an enterprise LLM - because it does not need to know poetry. It needs to know exactly which test to run when Workday renames a field, and how a change ripples across finance, HR and supply-chain processes. Argus reasons about workflows, troubleshoots errors, and orchestrates the platform's automation.

On top of Argus sits Opkey's agentic layer, launched in February 2025 as what the company describes as the first and only agentic AI ERP Lifecycle Optimization Platform. Rather than a single chatbot, it deploys a crew of virtual agents with job descriptions: a Configuration agent for mapping and migration, a Testing agent for automation and patch validation, a Training agent that generates end-user guides, and a Support agent for real-time help. Opkey pairs this with a "human-in-the-loop" philosophy - AI does the heavy lifting, humans keep judgment and control.

Who uses it, and how it sells

Opkey's customers are large enterprises with complex ERP estates - names referenced publicly include Gap, Pfizer, Cigna, Northwell Health, Myriad Genetics and The Wonderful Company. It reaches them two ways: directly, and through the systems integrators who run enterprise transformations, including KPMG, PwC and Huron Consulting Group. That partner motion is a quiet strategic advantage - the fastest path into a Fortune 1000 ERP program is often to make the consultants already in the room look good.

The business model is classic B2B SaaS: subscription access priced around reduced testing effort, faster releases, and lower risk during updates and migrations. Opkey markets outcomes like up to 85% less testing time and up to 50% lower ERP costs. Third-party estimates put annual revenue near $32.7M, though the company has not officially disclosed the figure.

Where it fits in the market

Opkey competes in a crowded test-automation field that includes Tricentis (Tosca), Worksoft and Panaya on the ERP-specific side, newer no-code players like Leapwork and Katalon, and the incumbents and open-source frameworks - Micro Focus UFT, Selenium, Playwright - that many enterprises still lean on. Opkey's differentiation is the combination of a purpose-built ERP AI model, a large certified test library, no-code accessibility, and an agentic layer that reframes the category from "test automation" to full ERP lifecycle optimization. Whether that reframing holds, the underlying problem is not going away: cloud applications will keep updating, and someone - or something - has to keep testing them.