How symplr lifted weekly output 44% while migrating 88 .NET projects to .NET 10

symplr is a leader in enterprise healthcare operations software and services, trusted by 9 out of 10 U.S. hospitals and more than 400 U.S. health plans. symplr helps healthcare organizations simplify operations through AI-powered solutions across provider data management, workforce, quality, and operational workflows. Behind these solutions is a distributed engineering organization that supports a broad suite of products across multiple technology stacks and generations of tooling.

symplr's engineering organization spans 18 teams building and maintaining a suite of healthcare operations products. Going into 2026, leadership wanted to make AI coding a standard part of how every team worked to lift overall velocity, modernize aging stacks, and grow capacity. The challenge wasn't adopting AI in a single team; it was establishing repeatable, AI-native workflows the whole engineering org could carry forward, and proving them out on real, high-stakes work rather than contrived examples.

One modernization workstream offered exactly that kind of proving ground: moving from .NET Framework 4.8 to .NET 10 across 88 C# projects. The codebase included Entity Framework EDMX (Entity Data Model XML) files that needed conversion to a code-first approach, alongside frameworks scattered across different generations of .NET tooling. Tackled project by project, the work would have taken months.

DevClarity partnered with symplr on a multi-month engagement, beginning with an intensive assessment to deeply understand challenges & opportunities across their worldwide organization. Foundational AI coding training then established Claude Code as the team's primary tool across the engineering organization, and a series of directed effort projects let teams apply what they learned to high-value use cases alongside DevClarity's AI engineers.

One delivery team turned to a focused workstream on .NET modernization and built the migration infrastructure they needed to move fast, not a one-time conversion, but a repeatable skill the team could apply, extend, and own:

• • Intensive assessment. Identified real challenges and opportunities within symplr's SDLC and set the stage for foundational training, replacing "random acts of AI" with a roadmap for genuine transformation.
• • Foundational AI coding training. Equipped the engineering organization with Claude Code workflows, repo-native context files (CLAUDE.md), and skill-building practices that carried through the rest of the engagement.
• • 8-step .NET migration skill. A dependency-aware workflow covering the full conversion lifecycle, from dependency graph generation and SDK-style .csproj conversion to NuGet package migration, API compatibility resolution, and build/test validation, with a tracking table to monitor progress across all 88 projects.
• • symplr-specific migration reference. A REFERENCE.md populated with the team's own dependency analysis, NuGet package mappings, API replacement tables, and Entity Framework 6 → Entity Framework Core guidance, turning general AI capability into symplr-accurate output from the first run.
• • EDMX-to-code-first skill. Built in under an hour, this skill converted 11 Entity Framework EDMX schema files into 389 POCO (plain old C# object) entity classes with proper data annotations, primary keys, and computed properties.

Between sessions, the team didn't wait for facilitation. They independently migrated roughly 40 projects, extended the skill, built migration tracking tables, and divided work across members. In parallel, five other delivery teams ran their own directed efforts, spanning Angular Ionic to React Native modernization, C# payroll rule migration to the cloud, OWASP vulnerability scanning across a legacy PHP codebase, Datomic to Postgres data migration tooling, and unit test generation paired with a reusable dependency-injection refactor. Each followed the same pattern: a real engineering bottleneck turned into a repeatable, team-owned skill.

By the end of the highlighted directed effort, all 88 projects were migrated to .NET 10 in approximately one week, building accurately with 0 compile errors. The team extended the workflow themselves, generated 389 entity classes with a skill built in a single session, and finished the full migration on their own after the second session.

Beyond the headline numbers, the durable outcomes were:

• • An 8-step migration skill that lives in the repo. Any engineer can run and improve it.
• • An EDMX-to-code-first skill ready to apply across all 8 Entity Model projects in the codebase.
• • Independent momentum. Between sessions the team extended the workflow, added tooling, divided work across members, and finished ahead of schedule without facilitation.
• • Human-in-the-loop checkpoints that caught edge cases, such as dual-targeting for controllers and EDMX cross-project dependencies, before they became blockers.
• • Repo-backed artifacts. Skills and context files are committed to the repo, owned by the team, and ready to share.

Across the full engagement, spanning six focused workstreams and symplr's entire engineering organization, engineers upskilled and built team-owned skills for modernization, security, data migration, and testing, reporting they were now +57% faster on repetitive tasks and had increased overall weekly output by 44%.

What symplr received. Org-wide AI coding training and upskilling that made Claude Code a standard tool across the engineering organization, plus six team-owned workflows committed to the repo: the 8-step .NET migration skill and EDMX-to-code-first skill above, plus five parallel efforts spanning frontend modernization, payroll rule migration, security scanning, data migration, and unit testing, each with its own symplr-specific reference documentation. The lasting value sits in repeatable workflows, repo-native documentation, and a team that kept extending the work on its own.