quality_assurance · finance · workflow
How Plaid grew AI coding adoption to over 75% of engineers
Plaid needed to shift hundreds of highly effective engineers to AI coding tools without stalling productivity, while navigating a fast-moving vendor landscape and the compliance constraints of operating in regulated consumer finance.
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 · Tool pilot evaluation
Plaid assesses AI coding tools by doing real development on large open-source or public projects to determine baseline quality before any procurement.
Tools used
CursorVS CodeJetBrainsSlackOktaLLMs
Outcome
Plaid grew regular AI coding tool use to over 75% of engineers, cut new tool pilot timelines from weeks to days, and ran a company-wide AI Day with 80%+ engineering participation and 90%+ CSAT.
What failed first
Simply announcing tool general availability did not sustain engineer adoption; after internal announcements, adoption quickly plateaued without dedicated ownership or follow-through.
Results
Time saveddays instead of weeks
Volume> 75%
Running sinceover the past six months
Grounding & classification
Source type: technical build writeup
27 fields verified against source quotes, 1 dropped as unverifiable.
agentic workflowcode generationcode diff prfailure mode describedhuman review describedmetric backednamed customerproduction runtime claimedtools describedworkflow describedfinancial servicessoftwarecycle time reductionemployee productivitytechnical build writeupquality assuranceagentic task execution