Stepping into a new leadership role often means inheriting risk you didn’t create. For most IT leaders, legacy applications are exactly that kind of risk: systems that still work well enough that nobody wants to touch them. Here’s how one CIO got ahead of this risk before it caught up to him.
He’d been in the CIO seat only a short time when he found what was waiting for him: four applications, all in production, all critical to daily operations, all end-of-life. Some were running on technology versions with no active security support remaining. The documentation on file didn’t reflect how the systems actually behaved anymore.
What the documentation didn’t capture lived somewhere else entirely. It lived in the memory of one person, the lead developer who’d built the systems years earlier and was weeks from retirement.
The Challenge
Most technology leaders in his position would have considered deferring this problem. But this wasn’t a risk he had the luxury of studying from a distance. It was already live, sitting under systems the business depended on every day, and he was the one accountable for it now, whether he’d had time to learn it or not.
What he needed wasn’t a long diagnosis. He needed someone who could move quickly and resolve the problem. The sooner it was resolved, the sooner he could move to more productive work.
”What we really needed was to reduce corporate risk — both security risk and the support risk of losing our key developer to retirement. With such a short timeframe, we had to find an accelerator to prove what was possible with today's modern tools.
ChrisCIO
The Guide
CODE Éxitos’ team of digital creators used the Athanor Method™ to lead the Assay and Transmutation of all four applications.
During the Assay phase, CODE Éxitos engineers analyzed the actual code and documentation, then worked directly with the people who used the systems every day. The team didn’t rely on a survey, but real working conversations. What features to keep. What code to retire. What parts of the application to modernize. The digital creator’s job was to pull out the knowledge sitting in people’s heads and turn it into documentation and a scope that could be built.
Deploying an agentic swarm (up to 110 agents working in parallel in this engagement) against the full codebase: reading source, tracing data flows, mapping dependencies, flagging anything undocumented. The raw pass on one application took about a week of compute time, and what came after mattered even more. Working from the structured findings produced in the Assay process, CODE Éxitos engineers then used three weeks to validate and stress-testing the findings. Experienced engineers separating live and critical information and code from what had simply piled up over the years. Automated and trained agents did the grunt work, but experienced people validated and approved the decisions to move forward.
”Instead of a single reviewer slogging through three codebases sequentially, our specialized agent swarm analyzed three different applications simultaneously. It’s like sending a team of precise auditors into a filing room—delivering structured, verifiable facts instead of guesswork.
Juan Carlos MendezCTO, CODE Éxitos
Two of the four applications were candidate for immediate Transmutation. They were fully modernized and redocumented with almost no additional effort after the thoroughness of the Assay. Upgraded, restructured, cleared of dead code, stripped of the personal data sitting in the original test fixtures and replaced with anonymized versions, and covered by automated tests for the first time in their history.
The other applications came through the Assay with something just as useful to our client: a Transmutation roadmap with defined effort, sequenced priorities, and real cost figures. Not a consulting deck, something he could actually act on.
”The agents showed us exactly how the code works; human leadership provided the organizational context to act on it. Blending AI speed with real business insight transformed an accurate technical inventory into a strategic roadmap that actually passes executive approval.
Juan Carlos MendezCTO, CODE Éxitos
Where It Stands Now
What a five-person team would have taken two months to do, a two-person team delivered in four weeks.
For our client, the outcome wasn’t just technical. It meant walking into a conversation with his leadership team with a real business win with AI: two modernized systems, two others mapped in enough detail to plan and budget against. He’d stopped managing a background risk and started making decisions in front of it.