Beyond the catalog: How Skai eliminated tech stack blind spots and achieved continuous governance
From manual spreadsheets to agentic automation: Bridging the gap between IDP visibility and infrastructure remediation at enterprise scale
The "Visibility-Action" Gap
Skai had already achieved a level of platform maturity many companies strive for. They had an Internal Developer Portal (IDP), a comprehensive service catalog, and automated "Gold Standard" maturity checks. However, they hit a wall: Visibility is not Governance. While the IDP could tell them what they owned, it couldn't tell them what was aging, what was drifting, or what required immediate action. This created three critical operational "friction points":
- The Spreadsheet Hassle: To manage deprecations, the team relied on a manual "Runtime Version Lifecycle" spreadsheet. Architects were forced to painstakingly track EOL dates for over 30 distinct technologies - from Kubernetes and MySQL to Spring Boot and Python packages.
- Deep Runtime Blind Spots: There was no automated way to see which specific runtimes were active inside microservices or Lambda Docker images. Identifying workloads relying on deprecated versions was a manual, error-prone scavenger hunt.
- Financial "Surprise" Costs: Without temporal awareness of when infrastructure would decay, Skai couldn't accurately predict AWS extended support charges. Budgeting for maintenance was reactive rather than strategic.
We had a catalog. We had maturity checks. What we didn’t have was a system that could tell us what was actually aging, what mattered most, and how to drive action across teams.
- Miki Manor, Director of Infrastructure Engineering, Skai
The Solution: Agentic Tech Stack Governance
Skai deployed Draftt as the "Intelligence Layer" sitting atop their existing infrastructure and IDP. By connecting directly to Skai’s AWS estate, Draftt began scanning EKS clusters, RDS instances, and Lambda functions to turn static data into Agentic Governance.
- ECR & EKS Intelligence: Draftt instantly mapped ECR repositories against live EKS deployments to reveal which microservices were running on deprecated or vulnerable runtimes.
- Drift & Decay Detection: The system highlighted underlying EC2 OS versions and tracked Pulumi-managed resources that had strayed from their declared state.
- Automatic Ownership Resolution: Draftt identified "zombie" infrastructure and resources with stale team tags, ensuring every issue was routed to a living, breathing owner.
- Waste Elimination: The platform systematically flagged idle resources and EBS volumes unaccessed for 90+ days, immediately impacting the bottom line.
The Impact: From "Ceremony" to "Continuous
By moving governance out of the meeting room and into an automated platform, Skai transformed their operational efficiency.
1. Predictable TCO
With Draftt’s predictive EOL tracking, the infrastructure team now forecasts AWS extended support charges months in advance. Maintenance is no longer a "surprise expense"—it is a planned part of the Total Cost of Ownership (TCO).
2. Risk-Free Infrastructure Upgrades
By gaining granular visibility into the EKS environment, Skai completely eliminated blind spots surrounding Kubernetes upgrades. This prevented unplanned disruptions and avoided emergency "fire drills" to meet vendor deadlines.
3. Governed Remediation
Instead of hunting through spreadsheets or losing tasks in a Jira backlog, engineers receive prioritized, owner-assigned issues routed directly into their daily workflows with integrated progress tracking.
We built a process around quarterly reviews to force teams to look at what mattered beyond daily feature work. Draftt turns that periodic discipline into a continuous governance system
The Bottom Line: For Skai, the partnership with Draftt represents the future of Platform Engineering: a world where the service catalog provides the structure, but the intelligence layer provides the action.
