- 97% have implemented AI, but 47% are waiting for meaningful AI ROI to be realized
- 79% say key operation decisions still require manual intervention
- AI could seriously help Q4 shopping habits
New research has claimed even though nearly all (97%) retailers have implemented AI in some form, more than two-thirds (69%) say they only respond to operational problems after those issues have already affected commercial performance, implying that most aren't planning ahead.
As a result, nearly half (47%) are still waiting to see measurable returns on investment from their AI spend, the report from UiPath found.
This comes as many struggle with the same challenges that have plagued AI adoption for years, with 42% still struggling with poor data visibility.
AI adoption is high, ROI is low
A third (35%) of UK retail leaders even continue to blame legacy technology for slowing them down, despite years of studies and reports implying that complex tech stacks and poor data quality are among the biggest blockers to successful AI. Delayed decision-making and inventory inaccuracies also contribute to delayed and unsuccessful AI rollouts.
Poor tech even extends to inefficiencies today, with four in five (79%) retailers saying most, almost all or all key operation decisions still require manual intervention, slowing response times and limiting AI's realistic impact.
"Often, businesses blame supply chain disruption when the real problem is that they’re making decisions with incomplete or outdated information," Retail Director Catherine Frame wrote.
UiPath says the companies that will see the most success with AI will be the ones who achieve "operational excellence," rather than the ones that blindly invest in AI. In other words, the same solid data and tech foundations that reports have been calling for for years.
Looking ahead, margin protection looks to be one of the biggest commercial risks as we enter the final quarter of the year, and the biggest one for commerce – while companies will struggle to achieve "operational excellence" and effective AI by then, it certainly highlights a major area where AI could support.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
