Study Finds AI Shopping Agents Erratic and Unreliable for Getting Best Deals
The Decoder · rss · 2026-08-28
Research from the Wharton School reveals that AI shopping agents are highly erratic and may not secure the best deals. A single external source like Wirecutter could shift product picks by up to 99 percentage points, and even changing the order of identical information altered outcomes. This suggests current AI agents are not yet ready for autonomous purchasing decisions.
More from Apps
- Google Search AI Mode can now book hotels and track flight prices — rmstein · 2026-08-28
- Iterating design elements with Figma Agent instead of one-shot generation — floguo · 2026-08-28
- Gemini Omni 1.1 integrates with Pika Labs, supporting video extension and 4K output — aziz4ai · 2026-08-28
- Claude Design's 'ask' tool UX criticized as annoying blind choice prompts — zeeg · 2026-08-28
- PortPal: A desktop dashboard to visualize and manage active ports — tom_doerr · 2026-08-28
- AI workflows: trading speed for multitasking efficiency — BenBajarin · 2026-08-28