Consumers are ready but human oversight remains key for AI—KPMG partner

Fast-food businesses are racing to get AI services right but accuracy remains a challenge.

Human oversight will be the key for fast-food businesses racing to get AI services done right, Emma Pitfield, Partner at KPMG said.

“With the democratisation of AI technologies, consumers are far ready for this type of technology especially if it provides them value. We’re seeing this obviously with KFC here in Australia. McDonald’s have tried it and although have discontinued a previous AI ordering trial have not ruled out using AI again in the future and I expect others to do so,” Pitfield said.

The analyst points out that whilst AI can streamline processes, particularly in order-taking, ensuring order accuracy and seamless customer experiences remains vital. "I don't think consumers are willing to accept that the order is not what they intended it to be. The consumer needs a seamless experience.” Pitfield said.

Challenges with order accuracy were reportedly the reason why McDonald’s had to abandon its trial use of AI voice ordering in drive-thrus in the US. Consumers posted videos of orders gone wrong, not recognising accents, and even adding things not on the menu.

In Australia, KFC is trialling a voice-ordering system in five of its restaurants. Kristi Woolrych, General Manager at KFC SOPAC, said that technology captures fluid speech in real-time and confirms requests via audio and visuals live with the customer, whilst also allowing customers to modify their orders with any upgrades and complex changes. If a problem occurs, a staff member is assigned to take over at any point.

Woolrych said they are currently evaluating the technology with staff and customer feedback before committing to building more in the future.

Unclear ROIs

Possibly one of the biggest roadblocks to the use of AI, whether it's for drive-thrus or other restaurant services, is how unclear the return on investments is.

Josef Chen, CEO and co-founder of UK F&B robotics and AI firm KAIKAKU, said not much automation and AI have been deployed in practice as most ‘retrofit’ solutions have failed to showcase a clear return on investment. 

“Karakuri, for example, had to file for bankruptcy in 2023 despite running pilots with Nando's. It is becoming increasingly clear that both AI and automation will require an end-to-end vertical integrated approach rather than trying to retrofit automation for a human-centric process,” Chen said

Pitfield said before investing in such technologies, it’s important for QSR brands to first identify what benefits are they after.

“I think QSRs should be trialling [AI technology] as they come to market but identifying the use case and the benefits are important. [QSRs should also] get feedback from both employees and customers in terms of the benefits. What are you after? Is it labour optimisation or freeing up your staff?” Pitfield said.

Pitfield said that the risk for early adopters is going to be the cost as well as ensuring franchisees understand the benefits.

For smaller chains, a wait-and-see strategy will make more sense as bigger chains with more R&D budgets go after various uses of AI. 

“A lot of that technology moves into a more white label, commoditised state as time goes by. There’s a hype associated with AI and a lot of immediate benefits as well but it’s just hone in on what are the proven use cases if you are a smaller chain,” Pitfield said.

Despite the challenges, the KPMG analyst predicts that within the next two to three years, AI will become widespread in the QSR industry, driven by the ongoing talent crunch.