Food Technology Magazine | Digital Exclusive
In the decade since founding Nuritas, your company has been able to compile and build out its library of peptides to harness AI to develop effective ingredients—two of which have come to market—PeptiStrong and PeptiYouth. What’s been one of the biggest “aha” moments for you in that time?
I think the biggest “aha” moment came when we were able to invent a technology that can find a molecule—specifically peptides—that would first have a benefit on a cell. It took years to develop the technology because you have to create the data for the AI. And then the next step for us was to show the effect on humans. Even today, no one else has taken an AI discovery all the way into human clinicals successfully and then out into the marketplace. We were the first to do that because we started early. That was a massive “aha” moment—that an AI platform can understand biology so well that it can find a molecule in nature that would have an impact in clinical. It’s huge. We’ve done about 17 to 20 double blind placebo clinicals so far with about 80% success rate. That is massive. That’s not happening without technology.
It’s about not only identifying a molecule that would work in clinical, but something that’s going to be stable in formulation, that’s going to taste okay, that’s going to be at a dose that makes sense, and that’s going to be orally available. So, it must be stable in digestion. All those factors are incremental to just efficacy in humans. With AI you can integrate all those factors so you can reduce the risk of something failing later on.
What are the next steps for Nuritas?
The next phase is really developing beyond the ingredients that we have today. We have a pipeline of ingredients in clinical today that haven’t hit the market yet. And we’ve learned a lot with our ingredients already in the market—including PeptiStrong and PeptiYouth. It’s now about bringing that learning back into all the ingredients we’re currently developing. And we are constantly working to understand consumer trends better so we can identify solutions for them. We’re doing a lot of that with AI. We are working to optimize the next ingredients. We have ones coming that address cognition, glucose, metabolism, and so forth. It’s a really interesting array of ingredients. We’re also now starting to incorporate with different technologies as well, such as wearables.
Ultimately an AI system allows us to create ingredients that are differentiated, that tick the boxes of what the large companies are looking for—maybe taste, cost, etc., but ultimately adds health benefits that can have massive impact globally. And that’s our goal: improving the lives of billions. That has been the goal of Nuritas from day one.
What advice would you give to food and beverage companies—either CPG or ingredients companies—that are interested in exploring how AI can help them, but don’t know where to start?
I think there’s two ways. I’ll start with the first one. It’s very hard to retrofit a completely new technology in an old traditional setting. It’s always easier to build from scratch. We were built initially with AI, so it’s a very different approach. But ultimately most groups today want to integrate AI for the right reasons because AI can enhance so many different aspects of a business. But what often happens that doesn’t end up working well is that a group of AI experts are brought in, put in a room, and expected to magically solve all the company’s problems. That never works.
It has to come from the top down and the bottom up as well. Everyone needs to be aligned in terms of what are the areas of the business that would profit most from having an AI system—it might be product development, marketing, formulation, whatever that area. Identify those areas that are key for the business that would profit from AI enhancement. That’s how you build—you start with the problem.
Then identify the types of data points that you create within that area. In formulation, for example, the problem with a lot of companies is that they don’t have records of all their work—what has worked and what has failed. The more data formatted in a way that AI can digest, the better. Record your data in a way that shows the decision making of what didn’t succeed, what went really well, etc., but in a fashion that AI can understand. You don’t need many people for that. Maybe just one person that is dedicated to that in one area of application—whether it’s marketing or a product. And just start building that data. That is already a good start. And then you can implement an AI system. And the difference that the AI can make with that data—in helping you decide better, not make the same mistakes, optimizing—is massive.
The second one is really identifying within your area other companies, whether from an ingredient perspective or other companies that use AI, and try to work with them. It takes time to develop collaboration with large companies, but ultimately it usually pays off.ft