Retail’s Agentic AI Shift is Coming Faster Than Systems Can Handle

Retail engineering teams have already made AI assistants part of their daily work. But full agentic AI is a different story. That’s according to global MIT Technology Review Insights survey of 300 senior technology executives. Our new industry analysis breaks down where agentic engineering in retail stands today and how fast that's about to change.

Inside, you'll find four insights:

• AI adoption is running at two speeds. 97% of retail engineering teams already use AI assistants for coding and testing every day — the highest rate of any industry surveyed. But only 27% have moved into agentic AI.

• Retail's next move is closer than it looks. 67% of retail leaders plan to deploy agentic AI within the next 6 to 12 months — the steepest projected adoption curve of any industry in the study.

• The obstacle is the stack. 63% cite connecting AI agents to existing commerce, POS, and ERP systems as their top barrier — more than double any other concern on the list.

• The payoff moves from tasks to the whole pipeline. End-to-end software development is set to become retailers' leading source of value by year three — up from just 20% today — as gains shift from isolated tasks to the full delivery pipeline.