Reducing Food Waste with AI in Commercial Kitchens

Hello and welcome to the latest edition of the Techstrong AI Leadership Insight series. I'm your host, Mike Vara. Today we're with Femen Gong, who's CEO for meta food X and they're applying AI to the way food gets managed in commercial restaurants and other kitchens with an eye towards reducing how much food we do waste, which is today considerable hangman, welcome to the show.

Thank you so much. Good morning. People have been talking about this issue for a long time and it seems like it's always been something that we could never quite solve.

So maybe walk us through here what you guys are doing with AI to kind of get us to the point where we are using the food that we have more efficiently and maybe not wasting nearly as much. Yeah, absolutely. And of course if you look at the big wine, uh, food waste could, a loss can happen right?

During, during the transportation from the farm all the way to the table. Then you have a big part is the food operation, right? And consumption and then all the way right to post-consumer.

Uh, and then when we look at this issue in the past there has been a lot of effort, great effort, that are very effective at creating the awareness about food waste. As you are fully aware, there are many solutions just to look at before the thing going to the trash bin, right? And that's where we have the most progress.

And today we really figured it out, says, okay, that is great. We have been doing recycling, repurposing the food and helping the environment. But then there is a huge part to tell us that we are doing it too late, too little.

And this is where, uh, to your question, I mean my background is in cybersecurity, right? I have been doing that for many years. One of the concept we learned there is this notion of, uh, look left right?

Or shift lab. I mean in this, in that case we are looking at this where is the root cause of the problem? What is the most effective way to prevent right threat and damage?

So in this case, applying to the food waste. And we realized that one of the biggest challenge for the food service operation is really lack of visibility from plan to consumption. And then we end up, we don't have the data to do accurate forecast planning.

And if you can do that, then what happened is you solve the efficiency, you have the eff efficient operation, but then food waste problem becomes solved as part of the prevention. So then almost like when you solve the problem, solve the infrastructure gives the operators the tools to actually put into place a lot of best practice they have actually accumulated in years. Uh, such as right small batch just in time cooking.

When they serve the food, they try to do first in, first out. It's interesting that they have all those, but they simply lack the tools to know where is the issue, right? How much they should make and where they should actually, how they can implement this best practice they have accumulated in years.

Because given the uh, face the issue with the labor, right? Um, the food cost, this is where if something takes a ton of effort for the operators to do manual work, uh, it's not gonna fly. So those are just quite a few things we learned and we put into our solution to help the operators.

Yeah. Yeah. So why do we lack the visibility?

'cause we've been ordering food forever in a day and people I guess have, you know, they have spreadsheets and all kinds of stuff to manage the process, but it sounds like some aspect of this thing is fundamentally flawed. So what is it? Yeah, and you absolutely and you know, it's uh, now in the head when you mention we have the spreadsheet, you know, soter, right?

Proven things and those indeed that's what people are trying to do today. But if you look at, come back to the visibility, right? If you use restaurant as a simpler case of a reference, what end up happening is people have been watching very diligently what they bought, right?

And that's their income inventory. And then some places they would use the POS actually to know how much they have sold. And when they talk about profiting loss, that's the two ends.

They put the numbers together, see what it is. And then if you now look at it in that process, there are actually a lot of blind spots. 'cause the issue is what you bought is the raw materials, the ingredients and what recipe you are serving.

How people are doing from the kitchen prep, trim, all the way to preparing the dish. What get consumed is actually the dish, right? One dish may have five different ingredients in them and that all become by assumption, right?

When people do today, when they do the purchase, they do rough guessing because that's the best way they can, they try to do a little bit inventory at the end of the day or some of them are doing like much less frequently. But this huge amount of guesswork is in there. Of course, if you scale that up to a commercial kitchen environment when you are serving right, thousands or several tens of thousand people daily, that's where the things adds up.

Uh, one of the biggest issue in that environment is simply overproduction. 'cause you don't have a better way of managing it than people go and buy experience than they have to earn on the site to make sure people do not have go without food, right? So that, that's the issue there.

Yeah. Mm-hmm. Do I also need to get smarter about the menu planning?

Because maybe I wanna put something on the menu that lines up with whatever I'm ordering in a way that uses that more efficiently. And I may not wanna have the same food on every meal, but at the same point, I, there's a lot of ways to cook chicken, right? Yep.

And, uh, this is absolutely a very important point the way, uh, with the data, right? We enable the staff and the team able to do typically two most important things. One is exactly the manual optimization because people do manual optimization all the time, of course, you know, both for flavor for change and also today, right?

There is stronger demand for eating healthy eating sustainably, uh, sustainably. So then that in increased a lot of, uh, chances or needs for many change, but you cannot do many change in the blind. So these where the data again come in because with the right solution, if you track the consumption day in, day out, based on manual cycle, you actually see what is the popular one?

What is this popular one, how the line up. So that's one of the critical input for a bit. And then another aspect just to add to it is actually, you know, day of the week, right?

And even during the day, you can imagine there is actually rush hours and peak and valleys. So those data that operator are using today before they don't have it. Now when they have that data, all of a sudden the just in time small batch processing become so much more effective that actually help both experience the quality of the food and also the waste.

Mm-hmm. And does that not also impact my staffing requirements? Because, um, different recipes and different times of day may require different numbers of people and different levels of expertise, but if I don't have visibility into that, I'll misalign the staff with the amount of food that's showing out the back door, right?

Yeah, you are absolutely right because those are two ways of that. One is if you do accurate forecast, accurate planning, and then, then you actually can schedule staff because a lot of staff in that environment is also on hourly staff, right? And so that, that's definitely one thing.

And university is actually interesting also because they leverage a lot of student for student both for as part of the education and helping out. So that's where this become like very, very relevant. Not only you can schedule ahead of time and reduce the amount of overproduction, so you do not need to use like the staff where you used to think you need.

And the second aspect is actually you can schedule the step out based on their hour. So both of them. And, and there is another added benefit, um, which we have got feedback from the operators who use it.

That is, uh, they actually realize it our solution, right? If you have a solution that is systematic, let's say you make the food consumption tracking very easy by staff following very simple instruction, then all of a sudden you find out you have very consistent complete data on consumption. At the same time you actually reduce the burden on the staff, on the staff with the children, uh, on staff training and so on and so forth.

So that actually is very interesting also helping with the, uh, labor challenge. A lot of folks will say, well, won't that just reduce the amount of food that, uh, theoretically these kitchens might be sharing with various charities 'cause they've overcooked or whatever. But, um, correct me if I'm wrong, most of that food doesn't make it there anyway just 'cause we don't know how to actually distribute it in the, the, the supply mechanics is doing that or too overwhelming as well.

Uh, yes, and I think both, on the one hand we give credit, right? There are folks that actually create that ecosystem for the food donation. Uh, indeed.

Um, although, you know, if you look at we fed right data, uh, unfortunately the participation rate is still low. The one of the reason cited there is exactly because one is the logistics because when the food reduction operators are already under the pressure for, you know, producing the food, serving and then that donation, so it's additional effort for them to do. And then there is also a second aspect because we do not have very clear, well-defined guidelines and safety requirement.

So, so that also put an additional burden. Now that said, uh, indeed there is room, right? Regardless what we do, um, there is always gonna be the food waste.

So the idea is by introducing the right solution, not only we track, let's see how much you prep, how much got served, how much is still sitting in the cooler right, than the warmer. And now with this degree of tracking, we actually not have a piece of exactly the kind of food where out, what condition and should we make the job right to connect that with the ecosystem that actually distribute that food or the needy and then that everything become more efficient. Yeah, absolutely.

Well, a lot of that wasted relates to food that wasn't even cooked yet. So that's part of the other side of the issue as well. Yeah.

Is this something that is applicable to commercial kitchens, a k organizations that are, I don't know, running kitchens for dorms or can I apply this all the way down to my local pizza shop or the local restaurants in my town where there's just a lot of turnover? Because a lot of those people, well, their first love is cooking and their second love or maybe even hatred is spreadsheets. Yeah, and uh, very good question because we started with let's say college and university because that actually presents the most challenging environment, the students, their case, their demand, uh, always kind in the frontier.

And then we have, we have gone, we have, uh, just last week we have, our first customer actually is Mayan, uh, resort. So we are very happy. And then we also have our first, uh, quick service restaurant.

And, and you, you mentioned restaurant because we know it's very important. There are of course from, you know, the quick service, like all the way to, uh, fine dining. They very, uh, they manage it slightly differently.

I would say with the, probably today, uh, in terms of product form directly, uh, the food dining, it might be the last one we'll go to the food dining does have need that they have communicated to us that is actually the food quality control at the exp uh, expedition. We actually, because our scanning, quick scanning technology, getting temperature and food condition, everything. So there is a use case there, but of course we need to, uh, study more on that.

But in the other cases where anytime we do have RA volume of the food moving through, they do use, uh, in addition to made to order, they do use, even if it's a very small batch, right? They are very flexible. They'll do small batch, some small batch, some are not.

This is where number one is that efficiency forecasting is still important because today this still cannot manually track right? Throughout the day, day of the week, how the traffic is going. That's one example that this data will have.

The other thing they actually care dearly is about the quality, about the freshness of the food, about when it's going out. And we have this first quick service restaurant customer. They actual leverage our system where when the food is going, coming out, when you scan, it takes a timestamp based on your need, based on the dish.

You say, okay, this need to be checked in 15 minutes or in 45 minutes, I want it to be pulled and replaced. So that, that is how it's becoming part of that infrastructure because they care so much about the quality. But if you devote right human, then human is creative, but they are not very mechanical.

That's where I think we, we can see the technology mesh. Uh, with the staffing. Well, yeah.

Last question my friend. You know, you worked in security prior to this and I think you're one of the early founders of Palo Alto Networks. It's a big jump from cybersecurity to food distribution and preparation.

How did that come about? Yes, and very good question. I would say if there is a little step up for me for tackling this, I believe it's a very highly impactful mission is because I did the first small step that was I went to a ride sharing company.

What attacked me there was the first step says, okay, daily on the platform you're talking about, you know, 10 million or 30 million of the people riding it. It's a transportation, it impact everyone. And there is of course a lot of issues to be dealt with.

That was my first step. I actually spent time there, learned about getting closer to something, uh, affect a mass amount of people versus right. IT and enterprise.

And then of course, the biggest impact for me, biggest motivation for me is I feel learned enough about the technology, seeing the technology applied differently. Although in cybersecurity, solving different problems, right from enterprise endpoint to the cloud, um, this is where I feel confident once I learn how bad the food waste problem is. That's really the beginning point.

The moment I learned, you know, we are talking about 300 to 400 billion a year just in us, that total loss, that translate to dollars because how we are managing our food supply and just felt, yes, and we can do something. And we actually now went ahead and proven it. So we are really happy seeing the impact, right?

Going to the customer and what they are seeing, what this helps them in terms of efficiency, in terms of then preventing the food waste. Yeah. All right folks, you heard it here.

AI and advanced analytics can make a massive impact on, uh, food wastage. But it's bigger issue than that. When you think about it, I mean, it goes all the way back to, um, how many cows do we need to have, how many pigs, how much, you know, vegetables and farming and all of that requires energy and all of that consumes money and creates carbon and has impacts on the planet.

So it all starts with one small thing, but it can have a big impact. Hey buddy, thanks for being on the show. Thank you so much, Mike.

All right. And thank you all for watching the latest episode of the Textron AI Leadership Insights series. You can find this episode and others on our website meeting, invite you to check them all out.

Until then, we'll see you next time.