Coinbase says AI is overhauling compliance — and cutting through repetitive work
Coinbase co-founder and CEO Brian Armstrong revealed on X that a recent AI-powered upgrade has produced “huge efficiency unlocks” across the exchange’s workflows, with particularly strong results in compliance systems.
Armstrong said the company has effectively rebuilt its workflows around the new models, reporting a dramatic decrease in restriction resolution times — a metric the CEO says has improved by roughly 90%. He added that AI is handling the repetitive heavy lifting while humans continue to validate outcomes to keep security tight and refine the models.
The comments follow a deeper dive from Coinbase’s VP of Product, Dor Levi, who last week emphasized that compliance at a crypto exchange is far more than simple sanctions-list checks. Levi argued that the bulk of compliance work requires interpretive judgment under uncertainty, and cautioned that merely automating existing procedures misses AI’s bigger potential. “Done carefully, with proper controls and human review, models can explore more context, test more hypotheses, and surface more inconsistencies than any single analyst could reasonably do case by case,” he said.
These efficiency gains come amid workforce reductions. As reported by Bitcoinist, Coinbase announced roughly 700 layoffs earlier this month — about 14% of its global staff. Armstrong attributed the cuts to two main drivers: a slowdown in the digital-asset market and the integration of AI into core operations. The company expects the layoffs to be largely complete by the end of Q2 2026.
Market context: Coinbase remains the world’s second-largest exchange by 24-hour spot trading volume, per CoinMarketCap, logging about $1.5 billion in the latest window versus Binance’s nearly $8.4 billion. At the time of writing, Bitcoin is trading near $77,200, down roughly 2.8% over the past week.
Bottom line: Coinbase is betting that AI — paired with human oversight — can speed and scale compliance work, freeing staff for higher-level tasks even as the technology reshapes headcount and operational structure.
Read more AI-generated news on: undefined/news
