Case study
UX Friction Analysis 10x Faster for a Silicon Valley Analytics Company
How SquareShift built a hybrid pipeline on BigQuery, Vertex AI, and Gemini that organizes billions of user events and finds software friction ten times faster than manual review.
Book a session →
Client
A Silicon Valley digital analytics company that helps enterprises find and reduce friction in their software.
Industry
Technology
Services
Stack
BigQuery · Vertex AI · Gemini
How SquareShift delivered it.
The challenge
The client's enterprise customers generated billions of recorded user actions, and manual review of that data was slow and gave inconsistent findings depending on which analyst was looking at it.
What we delivered
SquareShift built a hybrid AI pipeline on BigQuery, Vertex AI, and Gemini, combining rule-based checks with AI to organize the raw events into one scalable data layer instead of many separate streams.
The pipeline links errors to retries, finds labeling mistakes, and measures time lost against an ideal user journey, with the deterministic part of the pipeline giving the same repeatable finding regardless of which analyst looks.
The outcome
Friction analysis now runs 10 times faster than manual review, and hidden friction patterns are surfaced and prioritized automatically instead of missed.
The value of AI here isn't replacing the analyst's judgment — it's giving every analyst the same repeatable baseline to judge from, instead of each one reading the data differently.
AI & Machine Learning Practice Lead, SquareShift
Related reading: vulnerability chatbot on Vertex AI and HR chatbot on Vertex AI.
