The challenge
The vendor is moving enterprise customers from legacy, CRM-native products onto its own cloud-native platform. Each migration historically meant large teams doing highly manual work:
- Converting custom Apex business logic to C# .NET. One enterprise customer alone had roughly 2,500 files of custom code to migrate.
- Re-creating field mappings from a legacy contracts system by hand: roughly 400 mappings per client, across a portfolio of some 300 clients for a single end customer.
At that scale, manual migration is slow, error-prone and expensive. The vendor needed a partner who could bring AI tooling into the delivery process itself, not as a demo but as production migration machinery.
What we did
- Stood up a dedicated AI-enabled migration team with an intentionally elastic model: start with a small senior core, scale with the migration backlog
- Built an AI migration wrapper for the legacy-contracts-to-CLM migration path, delivered as a fixed-scope engagement: AI-assisted conversion tooling wrapped in repeatable process
- Applied AI-assisted code conversion to the Apex-to-C#/.NET workstream for a global health technology customer, using LLM tooling to translate, refactor and validate business logic at file scale
- Automated the legacy field-mapping process for a major law firm's contract migration, replacing per-client manual mapping with tool-generated mappings reviewed by engineers
- Codified the approach into a reusable migration playbook (built on Claude Skills), so the vendor's own engineering organization can run migrations independently over time
- Prototyped a customer-facing AI support experience on top of the vendor's internal AI platform, including the guardrails that keep internal content out of customer-facing responses
Results
- Migration capacity that scales with tooling rather than headcount
- A fixed-price AI migration offering the vendor can resell into its own customer base
- A documented, transferable playbook: the engagement is designed to make the client self-sufficient, not dependent
- Follow-on AI and data-platform workstreams won on the strength of delivery
Tech
Apex, C#/.NET, LWC, React, Next.js, PostgreSQL, Claude and Claude Skills, MCP