AI is not a human

But it is helpful! A Q&A with Rachael Nemeth about the (good!) potential of artificial intelligence

Lately, any news about artificial intelligence (AI) seems fascinating or terrifying, with little area in between. This week, the so-called “Godfather of AI” broke ties with Google, where he contributed work for years,  in order to speak out about the dangers of the technology he helped to develop. That doesn’t exactly inspire confidence. 

But plenty of others, including restaurant industry leaders, have already spoken about its promise. 

Rachael Nemeth is the co-founder and CEO of Opus Training. She holds an admittedly rosy view on the potential of artificial intelligence to shape both Opus and the future of hospitality, but that’s because she’s already seen how it can transform employer-employee interactions. 

Opus offers mobile-based training¹ for “deskless” employees, which make up 70 percent of American workers. Its customers are mainly restaurants and restaurant groups: The Smith, &pizza, Luke’s Lobster, Sugarfish, and many more. 

Nemeth said she sees a lot of promise in the technology as a way to streamline work and improve conditions (and costs!) for restaurants; not to replace their people. 

“There has to be a truth and an honesty around the fact that AI is not a human,” she told me in a recent interview. 

“I think we’re in the first phase of managers and operators learning how AI can benefit them and how they can use it,” Nemeth continued. “I don’t think it’s a steep learning curve, but I do think there’s a trust curve, if that term even exists. It’s about getting people to understand that this is not going to replace the need for you to come in and add a human element.” 

Here’s more of our recent conversation, where Nemeth explains why she’s looking forward to the benefits of a buzzy emerging tech with the power to transform the business. 

Our conversation has been edited for length and clarity. 

Expedite: I imagine you have a good understanding of how AI could affect your business. Have you been watching the space for a while? 

Rachel Nemeth: “Of course, AI isn’t new, but now it’s mainstream. We’ve always kept our eye out for pieces of technology that can help us help our customers save money, make more money, and do more with their people. When the OpenAI craze started happening, I think we were as excited as the next business because it enabled so much for us.” 

And your customers — restaurants — are excited about it, generally? 

“I think they're stoked about it. Everyone has a cautious optimism about AI, which is healthy. I think it's the way we should be viewing all forms of technology. It's being excited about the ways that it can enable us as humans and being thoughtful about ways that it can help us do our jobs better. But we need to be equally thoughtful about ways that it can be harmful to others.

“Right now most restaurant technology companies are leveraging AI for diners. Opus is helping employers, and that’s where I see opportunity. The industry is seeing the highest quit rate of the past three years. There's a labor shortage, and there's more need for efficient training. We're trying to fill up this bucket with holes in it, and you can't do that without putting some putty in to do it faster.”

I think that's an important distinction, because a lot of the pushback I hear is about how AI might harm the restaurant experience for diners. That's where I think some of the hesitancy is, because it's like, well, this isn't hospitable. It's inhuman, why would you do that in the restaurant business? But using AI in this way is very different, as you've said. 

“Exactly.” 

Can you talk through a current use case for Opus and AI?

“I can talk about two, and they’re both relevant. Since our founding days, we’ve been leveraging AI through our translation technology. We automatically translate every course, every message, every interaction that you put into Opus into 100 global languages. You’re inclusive, your operational efficiency goes up, but your operating costs go down. If I had a nickel for every time, a manager translates a message into Spanish, or a training manager translates a lesson, or an HR manager hires an interpreter to live to deliver sexual harassment prevention training, I’d never have to work a day in my life again; these are common interactions that people are spending hours translating. We built a proprietary machine learning model and can achieve 95 to 99 percent translation accuracy, which is better than most humans. 

“The second way is more recent, and it has to do with what we call the content problem. On the employee-facing side, restaurant managers are constantly building SOPs [standard operating procedures] and recipes and menus and pre-shift notes and policies — you name it. Our AI powered Content Builder helps you take that content that you have on paper and convert it into a lesson 20 times faster than legacy solutions. It takes your inputs and trends and transforms that into an interactive lesson, so that your team actually learns it and can demonstrate it on the job. It solves a big problem with training nowadays — if you throw a packet of information in front of someone and say, ‘memorize this,’ of course, they don't memorize it and can’t put it to use.

“We actually have numbers on this, from one of our restaurant partners, Just Salad. Before Opus, about 50 percent of manager time was spent in the office, doing admin work, like training, following up with people, XYZ. Now, it's about 10 percent of their time. 

“Just Salad has also seen frontline labor costs for new store openings go down by five to 15 percent. We've been able to show that restaurant openings can get cheaper when you can optimize training. And frankly, anyone who's been a part of a new store opening knows that there are last-minute changes to operating procedures, because the footprint changes from store to store. So the ability to be able to spit out a new lesson quickly and ship it to your team so they can complete it with the accuracy that you need within a shorter amount of time means that your labor costs go down. Plus, people have higher uptake. There's a win-win for both the manager and the frontline worker.” 

You sound pretty excited. 

“I think the next thing to be excited about, and what most restaurants should be excited about, is the fact that AI gets smarter the more you use it. The more we train the models, the better they can serve employees. The more you use our content builder, the more it picks up your tone of voice. Every restaurant culture has its own language, and it can learn that behavior. We're not pulling from the billions of data points on the internet. We're pulling from your existing working documents that you use every day, and we're just helping you make them better.

“I guess I do see a lot more of the sunny side of this.”

It’s not wrong to look at the sunny side, though, because all of this is true. 

“No robots are going to come out of your screen and eat you alive. We’ve been using spam filters and phrase recommenders and face ID for so long — that’s AI doing the work. But those are consumer-facing technologies. Now everyone has a piece of that AI and they can use it to do better, faster work.”

1. A previous version of this piece referred to Opus Training’s model as text-message based, which was true a few years ago — but is no longer the case. Opus has been app-based for two years.