AI transformation fails when imposed from the top

Adoption succeeds when it helps managers work faster and make better decisions.

AI adoption in quick-service restaurants tends to fail because it is driven from the top down rather than starting with the frontline staff who actually understand the problems, according to GOPIZZA founder and chief executive Jay Lim, who said during a webinar hosted by QSR Media.

At the AI-Powered Restaurant: Driving Site-Level Adoption at Scale webinar, Lim said too many operators approached AI the wrong way round, starting with the technology rather than the problem. "We always have to start from what problem do we want to solve," he said. "Because AI can do so many things, we tend to say, okay, let's solve everything through AI, and the management gets overwhelmed, and the frontline staff get angry."

He said imposing AI from above left frontline workers demotivated and fearful for their jobs. The solution, Lim suggested, was to narrow the focus to small, specific missions identified by frontline staff themselves.

Lim pointed to GOPIZZA's operation in India, its largest, where the company hires and replaces around 100 people every month, placing a significant training burden on its team. The training team had wanted a customised learning management system for years but could not previously justify the cost.

Rather than commissioning developers, a training manager with no prior technology or development experience listed every function and feature he wanted. The team turned that specification into a prompt and built the tool for him.

The impact, according to Lim, went beyond the company's bottom line. Because the solution had effectively been built by the staff who needed it, it shifted the mindset of the wider frontline workforce about what they were capable of. He said the approach was gradually transforming the organisation, to the point where his frontline staff now wanted more AI than he did.

Nick Palamaras, commercial account executive at Crunchtime, echoed the point that genuine adoption is measured by whether AI changes what happens inside the restaurant, not by whether a feature is switched on.

"If you still have managers spending hours building schedules, doing counts, chasing compliance tasks, digging through reports, then AI really hasn't been adopted," Palamaras said. "Real adoption to me means AI is helping managers be faster, make better decisions without any more work. If AI is delivering this, adoption will naturally follow."

As an example, Palamaras said one organisation shifted from manually counting inventory to relying on 95% of the suggested orders generated by Crunchtime's solution, saving significant time.

Emma Pitfield, partner at KPMG, agreed with Palamaras and drew on her experience working in both restaurants and headquarters. She stressed that adoption had to reach every site without operators running parallel legacy processes alongside the new technology.

“It's not real unless it's in the restaurants," Pitfield said.

Pitfield said building trust with restaurant managers was central to this. Good managers ran their restaurants using a mix of "art and science," so AI had to be introduced through a human-centred approach that complemented their judgement rather than overriding it. She added that change management investment was critical for organisations deploying AI technology.