Summary

This early field note captures Abi experimenting with ChatGPT as a way to explore common problems in online shopping and ecommerce journeys. Written in 2022, it asks what AI could already identify about why people do not buy online, why they abandon checkout, what frustrates them after purchase, and what makes an online shopping experience feel trustworthy and usable.

The themes are strikingly familiar: security concerns, uncertainty around personal and financial information, technical issues, slow websites, poor product information, unreliable shipping, weak customer service, and painful returns. The carousel also highlights the basics of a good ecommerce experience: clear product information, simple checkout, reliable delivery, helpful support, and transparent return policies.

Seen now, the piece is useful as an early marker of Abi’s interest in how AI could support UX and ecommerce thinking. It also foreshadows later Corpus themes: trust, clarity, expectation-setting, operational reality, and the gap between what customers are promised and what they actually experience.

Topics

• Early use of ChatGPT to explore ecommerce behaviour

• Why people hesitate to buy online

• Common reasons customers abandon checkout

• Post-purchase dissatisfaction around quality, delivery, support, and returns

• What makes an online shopping experience clearer and more trustworthy

• Early predictions around mobile, augmented reality, AI, and personalisation in ecommerce

Best for

UX researchers, CRO practitioners, ecommerce teams, product teams, digital strategists, and anyone interested in early practical uses of generative AI for customer experience thinking.

Background

This is a useful early artifact in Abi’s wider body of work because it shows the beginning of a thread that later becomes much more explicit: AI is not just a tool for generating copy or ideas. It can become part of how teams inspect customer problems, surface patterns, and think about trust, expectation, and experience design.

It also connects cleanly to Corpus because many of the issues surfaced here are not just interface problems. They are trust and reality problems: unclear policies, poor delivery expectations, weak support, damaged goods, and information gaps.