MCP apps: shipping outside the roadmap

A small team with permission to skip the process — shipping MCP-powered capabilities weeks ahead of any formal roadmap.

Role
Founder/sponsor of the effort
Team
Small cross-functional crew [VERIFY] size
Timeframe
2025–2026
Outcome
Capabilities shipped outside formal roadmap; fed back into platform strategy

The bottleneck

Enterprise roadmaps move at enterprise speed. The Model Context Protocol didn't. When MCP made it possible to put Workato's capabilities directly inside AI assistants, waiting for the formal process meant arriving after the moment. So I built a different kind of team: small, scrappy, empowered to iterate and ship without a PRD in sight.

How it ran

[VERIFY] team size, cadence, what shipped — list 2–3 concrete MCP apps. The rule was disposability: everything shipped was allowed to be wrong, cheap to replace, and instructive either way. The team's output fed the platform strategy — MCP evolved from an API-platform component to a first-class project asset with verified user access — but the deeper output was proof that a design-led team could ship production capability outside the machine.

The formal roadmap is how an enterprise ships what it already believes. The skunkworks is how it learns what to believe next.

Outcome

[VERIFY] apps shipped
Weeks, not quarters
MCP → first-class platform asset
[IMAGE] MCP app in an assistant surface