In April 2025, we released Imogen®, the first end-to-end, behavior-based mainframe rewrite platform. And what a year it’s been.

‍

‍

Proof, not promise

We've proven our thesis that behavior-based, test-driven modernization is both safer and faster than conventional approaches. Generative AI has made it practical — and cost-effective — to execute on the gold standard of mainframe migration: incrementally rewriting an exact functional and behavioral replica into modern code.

‍

Imogen extracts all the business rules - not just a summary - and creates modern code using them as a harness. Each rewritten slice makes change easy and non-disruptive, enabling continuous modernization rather than a single, high-stakes cutover. Customers are then able to rapidly modify functionality and architecture knowing they have a stable foundation, even if the old system was not well understood.

‍

This is what it looks like to rethink an approach in light of what AI makes possible — not simply to bolt AI onto traditional, broken migration models. While it was encouraging for Gartner to recognize us as a Gartner® Cool Vendor in AI Code Assistants last June, we’re especially pleased to share a sampling of our customer projects:

‍

• A critical payments system for a large North American bank, modernizing COBOL, Assembler, and Easytrieve to Java on AWS — with characterization testing and backtesting to prove data equivalence — as the bank navigates new regulatory pressure.
• A business-critical COBOL workload for a major automotive company, achieving full functional, performance, and integration equivalence within a highly locked-down, air-gapped environment. 
• An ageing extended warranty system for a global industrial manufacturer, completed in four months onto AWS with our partner Thoughtworks — an 80% acceleration over initial estimates. 
• A sales system for one of North America's largest retailers, a project that’s currently running well ahead of schedule.
• A critical claims system that drives 30% of mainframe consumption at one of Latin America's largest insurance companies. Watch the customer story from Google Cloud Next 2026.

‍

‍It takes an ecosystem

Our partner ecosystem continues to grow. As part of the effort to tackle the immense mainframe estate in the public sector, we were pleased to be added to the US Government’s GSA Schedule and Tradewinds Solutions Marketplace. Google and AWS also added Imogen to their Marketplace platforms. 

‍

With deep thanks to Thoughtworks and Leidos in the US Federal space for the pioneering work we’ve done together, we will be rolling out a formal partner program next quarter to support more delivery partners. 

‍

‍Product acceleration

We’re getting faster. Much faster. 

Acceleration is happening on two fronts: how fast Imogen can rewrite applications, and how quickly we can broaden Imogen’s capabilities.

Our test-driven approach harnesses a non-deterministic technology to produce deterministic, auditable outputs, eliminating the traditional long tail of debugging and testing. As of last month, we're consistently replicating upwards of 10,000 lines of COBOL per week per engineer, generating production-ready code proven equivalent against actual production data. In fact, we recently broke our record at a F100 North American retailer, completing a job with 11,578 lines of code in 28 hours, fully verified.

For a deeper look at how Imogen differs, we borrowed a concept from Vikas Pujar that calculates how long it would take to rewrite the whole universe of COBOL, and created a side-by-side comparison of our approach versus traditional ones.

On the second front, Imogen is now handling an expanding range of what’s actually in real mainframe environments: more legacy languages (COBOL, JCL, PL/I), more workload types (Batch, Transactions, Screens), and specialized workflow tooling that accounts for the wide range of complexities modernization teams actually encounter.  We’re also embracing the technology ecosystem, integrating with important discovery and analysis tools like Google's Mainframe Assessment Tool.

‍

‍Brave new workloads

Working with our customers, we're reminded that migrating mainframe systems remains an act of courage: that even with a safer, non-disruptive approach, the risk calculus shifts instead to the unfamiliar.  People naturally seek proof and the safety of the tribe. 

‍

We're now providing both confirmation and community: concrete evidence that production-ready code can be delivered in a fraction of the time required by conventional translation approaches (even those using AI), and a growing community of like-minded practitioners who share the vision that AI's greatest promise lies not in improving existing processes, but in reimagining them entirely. We invite you to join us.

‍