Optimising ingredient manufacturing through data modeling

Ask anyone who knows her personally or professionally; Riana Lynn, founder of Journey Foods, has always been a quantitative person. She likes to talk about data, especially when there is a high degree of certainty about percentages and probabilities. Combine this with her research background in genetics, food and chronic diseases, and it is hardly surprising that her career path took her rapidly towards developing a start-up that harnesses artificial intelligence to get better food to consumers faster.

After a two-year stint researching genetic data at the University of Chicago, followed by a year working in business and public engagement in the Obama White House, Lynn entered the start-up world early on in her career, launching an e-commerce juice bar brand called Peeled. Her career has focused on the intersection of food and technology ever since, but she has never been able to get one idea out of her head.

“On the research side, we’re suffering so much from either lack of food or the abundance of food throughout all populations. While on the other side, when it comes to food companies, challenges like logistics and regulatory hurdles often limit us,” she says. “I realised that the focus should be on finding ways to save companies time and money and rapidly decrease the timeline in getting better food to consumers. The only way to do that with the number of people we have to feed every day is with data.”

Apply AI to food product development

Lynn has always been passionate about trying to understand the broader ecosystem of ingredients globally. “If you run a food company, everything from ingredients to understanding nutrition is quite a hurdle. It’s quite tied up in the third-party juggernauts of the system. There’s not a lot of transparency, not a lot of great user experiences. [Journey Foods] started with wanting to know more about this business area and apply that to a food product,” she explains.

In 2019, she gave a presentation to 1,500 people in the industry about the potential for AI to accelerate food product development. “Everyone seemed interested in how they could apply AI. Some of the world’s longest-running food companies, which are over 100 years old, were asking me about this and so I knew something was there,” she recalls.

But she did not want to simply provide consultancy for these companies. Lynn thought it would have a much greater impact to give them a tool they could apply internally. She spent the next six months building a platform to connect data to AI and machine learning models. The result was a piece of software, the Journey Foods platform, based around data modelling and success models that she hopes is going to reshape traditional approaches to food.

An efficient middleman

Journey Foods’ software uses machine learning, artificial intelligence, data scraping and cohort analysis to recommend the most nutritious and sustainable ingredients for food companies.

“When we first started, we were just focused on nutrition and we added sustainability markers. But the most important point here – my goal – is to create the most actionable database in food,” Lynn says.

By analysing over 260 characteristics including general nutrition, mass macronutrient values and proprietary sustainability scores in its recommendation engine, Journey Foods not only categorises and analyses ingredients but also connects food companies with suppliers, acting as an efficient middleman in the supply chain.

This allows smaller ingredient companies, which have developed innovative systems but do not have the financial or marketing clout to promote themselves, to expand their reach. “We like to look at companies that are using new processes as well as better ingredients,” Lynn explains.

“For example, they might use current ingredients but create less waste or use less water or starch or have more nutrient density. These companies exist everywhere. They exist in the middle of the country in the US, they exist in South America, Portugal and West Africa and we need to make sure to connect them because they have creative, passionate scientists but they don’t have the billion-dollar budget of big ingredient companies to push both marketing and research.”

Save time and money

Lynn’s experience in the food industry showed her that, in many companies, you would have food scientists on one side and business experts on the other and they do not tend to mix much.

“Food scientists often have no idea about cost and supply chain and other predictive models that feed into the ROI of making these big business changes that create more sustainability or health,” she notes. “Cost predictions and processing predictions have to be a central core to these decisions, so we are very focused on ramping up cost data in ROI models for our customers.”

Since implementing the Journey Foods platform, Lynn’s customers, which include the likes of Ingredion and Unilever, have seen huge savings in both time and money, in large part because the platform helps them reduce the number of trial production runs required to switch to a new product or formulation.

“Sometimes smaller runs can take several days, weeks or months to get a trial out. But we’ve seen a reduction sometimes from 30 trials to four for changing over to a gluten-free cookie or an alternative like that,” Lynn says.

Getting to production faster also means savings in the hiring process. “That’s been our first goal; how do we decrease the time and money it takes to get a new formulation to the market and get products on the market for consumers faster,” Lynn says. “Those savings can then be handed down to anything from more expensive ingredients to reinvesting in better packaging.”

Becoming the best digital twin

Journey Foods’ client base is currently spread across North America, Europe and Asia. Lynn’s aim now is to continue to grow in southern Europe and expand to West Africa and South America, while also improving the efficacy of the Journey Foods database.

“We really want to focus on how to be the best digital twin to a lot of processes that exist today,” she explains. “A big goal of ours has been to extend out the integrations that we have and partnerships we have across the world so we can solve a lot of issues around collaboration.”

For example, Lynn is looking into the other cost tools that are out there and whether there are nutritionists that could use the Journey Foods tool to help their clients. “Are there universities and food science programmes that can apply our data and tools?” she muses. “How can we continue to extend out the data in various industries and career types to continue to drive impact within food?”

At the same time, Lynn and her team are continuing to hone their AI and machine learning models. “We’re looking at everything from what’s the best packaging for a food product to what’s the best ingredient to lower water use. Instead of going through multiple trials and calling multiple distributors, how can we get that into a platform that tells food companies that information instantly? These are the questions that we continue to ask as we continue to work on our Journey Foods platform,” she concludes.

Increase efficiency using AI and bioinformatics

The Journey Foods platform is just one example of how AI can be used to improve the efficiency of the ingredient manufacturing sector.

Scientists with the USDA Agricultural Research Service’s (ARS) Western Human Nutrition Research Center (WHNRC), at the University of California (UC), Davis, have joined forces with over 40 researchers from six organisations to form an institute with a similar goal to Lynn’s: meeting growing demands in the food supply chain by increasing efficiencies using AI and bioinformatics. Their remit is slightly wider, covering the entire food system – from growing crops through to consumption.

The team, led by UC Davis, also includes UC Berkeley, Cornell University and the University of Illinois at Urbana-Champaign. The project is funded by a $20m grant from USDA’s National Institute of Food and Agriculture. “The AI Institute for Next Generation Food Systems (AIFS) is dedicated to accelerating the use of artificial intelligence to optimally produce, process and distribute safe and nutritious food,” says Dr Danielle Lemay, a USDA research molecular biologist at WHNRC.

In the area of food processing and distribution, which is the most complex aspect in the food system chain according to AIFS, the organisation will focus on enabling better outcomes in food safety, nutritious value and reducing waste through advancements in AI and process innovation.

To address food safety challenges, such as microbial and chemical contamination, AIFS is developing AI models to integrate food microbial ecology, chemometric and physical data sets, as well as creating digital twin models of food processing operations to include sanitation and food handling and transport to simulate pathogen transfer between humans and their environment.

Meanwhile, as demand grows to reduce chemical inputs and energy and water consumption during the production process, AIFS also plans to work on developing innovative food processing operations with a smaller production footprint.

This will include developing AI models to optimise food processing inputs (such as energy and water) and predict food processing outputs, integrating data sets from mechanical, thermal and chemical inputs. The organisation also hopes to predict and optimise product quality outputs such as texture, colour, flavour and nutritional value by developing digital twin models of food processing operations.