Ashley Stallings at Salsify made a call most CS leaders won't: when their Gainsight renewal came up, they didn't shop for a replacement. They built one.

In two months, her team migrated off Gainsight and into an internal platform called VIP (Value Intelligence Platform), built on Replit. It runs a Salsify-built LLM that pulls product usage, industry data, case studies, and pitch context into a single account view for every CSM.
This is what a well-governed, engineering-backed AI build actually looks like in production -- including the parts that didn't go smoothly.

Topics covered:

Why Salsify replaced Gainsight in two months and what forced the decision
How VIP was built on Replit and what it actually does for CSMs
The internal governance model: how AI projects get submitted, reviewed, and prioritized
Build vs. buy: the real argument CSP vendors make and why it's a false choice
What "handing it to a qualified team" actually means -- CICD, VPNs, data access controls
Health score problems: why they rebuilt from scratch and moved to segment and industry-specific scoring
Variance data as the new leading indicator -- top movers, not static red/yellow/green
The CSM cockpit vision: telling reps where to focus instead of letting them default to whoever's loudest
What they lost during the migration and how they're getting it back
The hackathon that produced a cohort-based outreach tool -- built by CSMs, not engineers

customer success, AI customer success, Gainsight alternative, build vs buy, CSP replacement, Salsify, post-sales AI, customer success platform, CS operations, AI in SaaS, HubSpot customer success, health score, CSM tools, post-sales operations, Infinite Renewals

Jeff Kushmerek is the founder of Infinite Renewals and the author of Retention Starts in Implementation.

🎙️ AI Customer Success Brief is sponsored by Infinite Renewals.
📩 Find Jeff on LinkedIn: linkedin.com/in/jeffkushmerek
🌐 infiniterenewals.com

Episode 1: AI in Post-Sales — Prototyping to Production

Most AI in customer success never makes it off someone's laptop. This episode is about why, and what it actually takes to get something running in production.

Jeff covers the mistake most CS teams make when they start thinking about AI, what the customer lifecycle looks like when you map it to real pain points, and the fork most operators don't know they're at: when MCP is enough vs. when you're building an actual application whether you planned to or not.

Real example included: a prototype built in Claude that became a containerized app running on Azure behind Tailscale VPN and Entra SSO — because the data lived in two systems that no MCP could bridge.

No metrics yet for most of this. That's fine. Sometimes the win is just getting parity — your internal systems wired together, working the way your team actually needs them to.

What's covered:

• Why starting with the tool is the wrong move
• Mapping AI to the customer lifecycle: onboarding, adoption, renewal, expansion
• The MCP path vs. the enterprise app path
• Real deployment architecture: GitHub Actions, Azure, Entra SSO, Tailscale VPN
• Why Intercom's help desk is a problem and what we built around it

About the show:
The AI Customer Success Brief is for operators past the prototype stage. Real implementations, real deployment paths, no vendor pitches. Hosted by Jeff Kushmerek, founder of Infinite Renewals. Sponsored by Infinite Renewals.