Your software

What "AI-ready" looks like in this area

• Your codebase has specifications AI can read — not just code AI can pattern-match against.
• Behavior is expressed in tests, not just in informal product knowledge.
• An AI tool changing the codebase can be evaluated against specifications, not just against compiler success.

What we typically find on Day 1

• Legacy codebases that work, but whose behavior is documented only through tests written years ago.
• Business rules embedded in code with no specification document — the code IS the spec.
• Customizations on top of vendor products that nobody on the current team understands.

What changes during a Business Transformation

• We write specifications for the parts of the codebase the transformation touches — what the code does, what it must continue to do, what it should not do.
• AI tools work against the specifications, not against vague intent.
• The specifications become artifacts your team owns afterwards — useful far beyond the AI work.

What stays the same

• Greenfield rewrites are rare. Most software gets specified and instrumented in place, not replaced.
• Vendor products stay where they are working. We do not propose replacing systems that are not the bottleneck.
• What changes is the specification layer around the software, not the software itself.

How this area connects to the rest

This is one of the seven areas an Enterprise Readiness review examines. The reference architecture that shows how all seven fit together is the agentic stack. The discipline that makes every engagement run the same loop is the repeatable process.