Back office ops · Production

Pinterest safely rolls out GitHub Copilot to all engineers, reaching 35% adoption in under 6 months

The problem

Pinterest initially prohibited LLM use pending legal and security evaluation, while engineers were eager to adopt AI-assisted development after using it on personal projects.

Workflow diagram · grounded in source
1
Engineer interest triggers evaluation
trigger
“many engineers expressed interest in adopting AI-assisted development and began using it for personal projects on the side, making them eager to use it at work as well”
2
Trial with 200 developers
validation
“we decided to run our trial with around 200 developers over a longer duration”
3
Copilot AI code suggestions
ai_action
“Copilot was particularly useful when I had to make a change in Scala, a language I am not familiar with. Being familiar enough with general language concepts, I could let Copilot take care of the syntax and still feel confident that I un…”
4
Security vulnerability scanning
validation
“Our security team leveraged vulnerability scanning tools to continuously audit all code introduced by both Copilot participants and non-participants”
5
Weekly NPS feedback via Slack
feedback_loop
“Qualitative sentiment feedback was collected weekly through a short slack bot based survey”
6
GA rollout to all engineering
output
“to increase Copilot adoption we ran training sessions, streamlined the process to get access to Copilot through integration into our access control and provisioning systems, and partnered with our platform teams to help folks understand …”
Reported outcome

Pinterest unlocked AI-assisted development in less than 6 months, increased user adoption by 150% in 2 months, and reached 35% of their total developer population using Copilot regularly, with an NPS of 75 during the trial.

Reported metrics
User adoption increase150%
developer population using Copilot regularly35%
time from idea to GAless than 6 months
trial NPS score75
Show all 5 reported metrics
user adoption increase150%
developer population using Copilot regularly35%
time from idea to GAless than 6 months
trial NPS score75
trial cohort sizearound 200 developers
Reported stack
GitHub CopilotslackvscodeJetBrains IDEs
Source
https://medium.com/pinterest-engineering/unlocking-ai-assisted-development-safely-from-idea-to-ga-4d68679161ef
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Pinterest unlocked AI-assisted development in less than 6 months, increased user adoption by 150% in 2 months, and reached 35% of their total developer population using Copilot regularly, with an NPS of 75 during the…

What tools did this team use?

GitHub Copilot, slack, vscode, JetBrains IDEs.

What results were reported?

User adoption increase: 150%; developer population using Copilot regularly: 35%; time from idea to GA: less than 6 months; trial NPS score: 75 (source-reported, not independently verified).

How is this back office ops AI workflow structured?

Engineer interest triggers evaluation → Trial with 200 developers → Copilot AI code suggestions → Security vulnerability scanning → Weekly NPS feedback via Slack → GA rollout to all engineering.

Related back office ops cases