it_support · saas · workflow
Pinterest builds a production MCP ecosystem with 66,000 monthly invocations saving 7,000 hours per month
Pinterest needed a unified substrate for AI agents to access internal tools and data sources, replacing bespoke one-off integrations for every model and tool. Additionally, spinning up new MCP servers required excessive operational work before teams could write any business logic.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · User initiates via surface
A user interacts with a surface like the web AI chat interface, an IDE plugin, or an AI bot.
Tools used
Model Context Protocol (MCP)PrestoSparkAirflowEnvoySPIFFE
Outcome
Pinterest's MCP ecosystem reached 66,000 invocations per month across 844 monthly active users, saving an estimated 7,000 hours per month for engineers.
Results
Time saved66,000 invocations per month
Volume844 monthly active users
Grounding & classification
Source type: technical build writeup
28 fields verified against source quotes.
agentic workflowai agentknowledge searchsummarizationknowledge basebuilder submittedhuman review describedmetric backednamed customerproduction runtime claimedtools describedworkflow describedsoftwareemployee productivitythroughput increasetime savedtechnical build writeupback office opsit supportagentic task executionai draft human approval