health technology

A custom CRM,
built to run the business.

Launch Labs partnered with Baden Sports to build, secure, and scale a custom CRM and operating platform — sales pipeline, ERP integrations, AI-assisted workflows, and the architecture to keep building on it.

THE CHALLENGE

A feature-rich CRM with the typical issues that come with AI-assisted code at scale.

A vibe-coded CRM, impressively complete on the surface — sales pipeline, email campaigns, lead gen, quote-to-cash, and live integrations to Microsoft Business Central, Microsoft Graph, Microsoft Fabric, SharePoint, RingCentral, and Apollo. The feature set was strong. But underneath, the build had the patterns AI-assisted codebases tend to fall into: critical security vulnerabilities, inconsistent access control across the data layer, and a codebase that lived in a single developer's IDE with no path for the rest of the team to safely contribute or extend.

Day-to-day, the company was running on something only one person could change. Every new feature came with risk. The system that ran their sales motion couldn't yet scale to the team that needed to work in it.

THE APPROACH

We rebuilt the foundation while keeping the CRM running — so Baden never stopped moving.

Five interlocking workstreams turned the platform from a single-developer prototype into a multi-team production platform.

"Launch Labs helped us build a CRM and operating system that'll grow with us. They secured it, built the architecture to scale, and set us up so anyone on the team can ship — not just me. We're moving faster than we were before, with the guardrails to back it up."

Casey Schindler

COO, Baden Sports

What Was Built

Three tools.
Zero guesswork.

Audit and three-bucket triage.

A pen-test-level review of the codebase produced a prioritized plan: stop-the-bleeding fixes for the immediate exposures, short-term hardening for systemic gaps, and long-term architecture work. The plan got real engineering on the dangerous half of the audit while keeping Baden's team focused on the work they enjoy.

A new architectural foundation.

Centralized authentication and authorization, database-backed sessions, normalized permissions, repository-level access enforcement, impersonation handling, and audit logging. One controlled layer for protected data access instead of scattered checks across dozens of server actions.

A wrapper pattern instead of a rewrite.

The old patterns got wrapped, not torn out. That let the new architecture take hold while the existing system stayed live — and let the build process enforce the new patterns on every new feature, so the codebase couldn't quietly drift back.

A real release pipeline.

The app moved out of one developer's IDE into a proper test environment, with a staging-to-production rollout the team could own. GitHub collaboration, AI-assisted code review, and a release cadence that let them ship without praying.

An AI enablement layer on top.

Hands-on training for Baden's team on Claude and Cowork, vibe-code project planning to keep future builds from repeating the same patterns, and custom Claude skills mapped to Baden's actual workflows. The platform stopped being a one-time deliverable and started compounding.

the result

1,500+
‍hours back.

Designer hours saved — time that went back to the work that actually requires human creativity.

MORE WORK

Have a workflow that needs automation?

Most of our clients starts with a conversation. We'd love to connect.

We respond to every inquiry within one business day.

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.

A custom CRM,built to run the business.

Three tools.Zero guesswork.

Read more stories

Have a workflow that needs automation?

A custom CRM,
built to run the business.

Three tools.
Zero guesswork.