For decades, modernizing enterprise systems has been one of the most expensive and thankless jobs in technology. That calculus is starting to shift.
Kabir Nagrecha grew up in the room where enterprise transformation happened. By the time he was eight, he was watching hundred-person teams navigate the grinding complexity of migrations that ran for years, cost hundreds of millions of dollars, and still went sideways more often than not. He watched consultants come and go, even as project scopes and timelines constantly changed. He heard the frustration of executives who knew their systems were aging and felt powerless to do much about it.
Two decades later, Nagrecha has a PhD in AI systems, experience at Meta, Netflix, and Apple, and a company built around the conviction that the technology to do this differently has finally arrived. He may be right.
A problem too big to keep deferring
The scale of the ERP modernization challenge is difficult to overstate. According to Gartner, global IT services spend runs at approximately $1.8 trillion annually, with enterprise application integration accounting for a substantial portion. The largest SAP migrations at Fortune 500 companies take three to five years and cost anywhere from $100 million to $500 million. And according to industry research, roughly 70% of large transformation programs fail to deliver what they set out to accomplish.
For years, the rational response for most CIOs has been to defer. Patch what can be patched, pay the extended maintenance fees, and hope the timeline can be pushed another year. But with SAP ending standard support for ECC in 2027 and roughly 10,000 ECC environments globally still needing to migrate, deferral is no longer a strategy. It is just a slower version of the same problem.
What has sharpened the urgency further is AI. Boards and CEOs are no longer treating AI investment as a future priority. They want returns now, and aging ERP environments are standing directly in the way. Fragmented data, inconsistent schemas, and years of undocumented customizations make it harder to deploy AI responsibly, harder to surface reliable data for decision making, and harder to move at the pace modern business demands.
According to Info-Tech Research Group’s CIO Priorities 2026 report, CIO success this year depends on disciplined value delivery, proactive risk management, and stronger data foundations. Legacy ERP environments undermine all three. According to Gartner’s 2026 CIO and Technology Executive Survey, 57% of CIOs are under pressure to improve productivity and 52% are under pressure to cut costs. The question most CIOs are hearing from their boards is some version of the same thing: when does the AI investment start showing up in the business?
For many organizations, the honest answer is not until the underlying systems are capable of supporting it.
Why the old model has lasted this long
The systems integrator model has dominated enterprise transformation for decades not because it works especially well, but because for a long time there was simply no alternative. The complexity of a typical legacy ERP environment, accumulated over twenty years of customization, undocumented business logic, and code written by people who left the company long ago, required a kind of contextual judgment that software could not replicate. So a global industry grew up to do it manually, and the economics worked out beautifully for the firms at the top, if not for the CIOs writing the checks.
That is changing. According to IDC, CIOs who prioritize tackling technical debt will be better positioned to adopt new technologies and prepare for more complex AI-driven transformation. Modern AI can now read decades of custom enterprise code and understand what business problem it was written to solve. It can map fields across inconsistent schemas, infer business logic when documentation is absent, and orchestrate the kind of complex, high-stakes cutovers that have historically required armies of consultants on-site for months. The work that made this problem so resistant to automation for thirty years is precisely the work that AI has just gotten good at.
Why no one has solved this until now
A new class of AI-native platforms has begun moving into this space, and most of them are taking aim at specific parts of the transformation process in isolation: accelerating data migration here, automating code conversion there, streamlining testing somewhere else. For CIOs managing complex, multi-system environments, assembling and coordinating those point solutions introduces its own layer of cost, risk, and program management overhead. The overall transformation still lives in human hands.
What the market has lacked until now is a platform capable of handling the full transformation lifecycle under one roof, from initial environment assessment through code migration, testing, data validation, and cutover, with the governance and audit trail that enterprise compliance teams require.
Tessera Labs is the first company to take that on at enterprise scale. Founded by Nagrecha and Ming Chang and backed by $60 million led by Andreessen Horowitz, the company deploys autonomous AI agents trained to manage ERP transformations end to end. Where a traditional program might require sixty consultants working over two years, Tessera says the same work can be done with a six-person team in a fraction of the time. Merck and Xerox are among the Fortune 500 companies already using the platform.
“Every CIO complains about this and hates it, but nobody in the outside world knows how much money companies spend on it,” said Seema Amble, the Andreessen Horowitz partner who led the investment. “Tessera’s platform makes transformation faster, more predictable, and continuous. That changes the economics in a real way.”
The platform is vendor-agnostic, which matters for organizations running environments that span SAP, Oracle, Workday, and a dozen other systems. And unlike point solutions that address individual workflow steps, it is designed to maintain continuity, governance, and traceability across the entire transformation process.
What comes next
The opportunity extends well beyond SAP migrations. Post-acquisition integrations, compliance overhauls, subsidiary spin-outs, and territory reorganizations all generate transformation work that currently consumes months or years of planning and execution. According to IDC, organizations that move beyond AI experimentation into enterprise-wide deployment stand to capture significant competitive advantages. Getting the underlying systems ready to support that shift is increasingly the prerequisite, not the afterthought.
With the 2027 deadline pressing and the backlog of unmigrated environments growing, the window for a measured, well-planned approach is narrowing. The CIOs who move first will have an operational advantage that compounds over time. The ones who wait will find the timeline, the cost, and the risk only getting harder to manage.
The technology to do this differently is here. The question now is who moves first.
