The news: Starbucks abandoned its AI inventory tool nine months after rolling it out to North American stores, Reuters reported. The tool made frequent errors, including mistaking different types of milk and miscounting items—making it more of a hindrance than a help to employees.

The big picture: Starbucks is not the only consumer-facing company reevaluating its AI investments, particularly as rising cost pressures—from computing expenses to elevated fuel prices—force businesses to adjust their spending.

• For Target, changes to AI pricing—particularly the shift away from subscriptions to a token-based model—are forcing it to adjust its strategy, Target in India president Andrea Zimmerman said at a Reuters conference.
• Uber’s COO Andrew Macdonald said that increased AI investments have not yielded increases in productivity or a notable rise in useful consumer features, per Business Insider.

Ultimately, most companies are reporting mixed returns on their AI investments. Roughly 2 in 5 private companies (43%) have only seen financial ROI in select use cases, compared with 39% who have seen moderate or significant returns across the board, according to a March Deloitte survey.

Implications for retail: As resources tighten, retailers will need to make tough choices about which AI investments to prioritize. Ultimately, the biggest priority should be those that move the needle now—whether by improving the customer experience, simplifying employees’ workflows, or increasing overall productivity. What companies should not do is pursue AI investments for the sake of gratifying investors, as that could inflate expectations and invite greater scrutiny if projects fail to deliver gains.