AI in nutraceuticals: how machine learning can optimise new product development

Artificial intelligence (AI) has burst onto the scene in recent years with a significant proportion of manufacturing and product-based industries now using it to optimise their workflows. This trend can certainly be seen in the nutraceutical space, as ingredient developers, supplement providers, marketers and quality testers have all begun to jump on the bandwagon to improve their operational functionality.

Because AI technologies are highly flexible, they can be used in a range of contexts. Therefore, companies should determine where this technology may benefit their business and how it might be incorporated to get the most out of it. To discover more about how AI can optimise the day-to-day activities of those working in the nutraceutical industry, Annabel Kartal-Allen spoke to PIPA’s VP of Science & Customer Success Catarina Abreu, and LEHVOSS Nutrition’s Head of Brands and Marketing, Teresita Rudà.

The nutraceutical market is highly saturated, so companies must push the boat out more than ever in terms of quality and formulation to create standout products. One way they can do this is by developing specialist or coactive ingredients that target popular health aspects, including cognition, gut and immune health. “AI — machine learning technologies, generative AI and natural language processing — helps us to analyse vast datasets of ingredients and their properties, as well as their bioactivity, molecular structure and any potential synergistic effects,” explains Catarina.

She continues: “This allows us to identify patterns and associations that we would otherwise miss if we were mining and analysing information manually. It can also predict which of these compounds will perform better, allowing researchers to identify promising ingredient combinations without extensive laboratory testing.”

Teresita believes that AI can also be a great time-saving tool in this context: “When combining ingredients into a nutraceutical formulation, the chosen extracts must be carefully assessed — which often requires a lot of investment in terms of research and time. Utilising AI in this context allows a company to filter out a significant amount of work, while working more efficiently. This can help businesses to quickly understand what ingredients work better together; and, from there, more specific research can be done.”

Although AI can be a highly useful tool for nutraceutical developers, Teresita warns that “there will always be a point when humans will need to determine what’s best.” Essentially, identifying plant extracts or active ingredients from a huge database can advise nutraceutical developers in terms of which way they should take their research … and what they should be focusing on to best cater to their target audience.

As well as facilitating the development of unique and personalised nutraceutical formulations, AI can help researchers to produce novel and more innovative flavours for their supplements. “AI can improve the process efficiency of flavour development, potentially cutting the typical 50–150 iterations down to much fewer,” states Catarina. “By making the flavour development process less based on trial and error and more systematic, companies can tap into the consumer market via AI-collated preference data to provide tailored flavour recommendations for specific demographics.”

“Flavouring is very subjective,” Teresita adds, “but certain combinations will generally appeal to some populations more than others. AI can support the process of coming up with unique flavour combinations — possibly offering more than a team of developers might think of — all while considering the trends seen in specific consumer groups. It can also help product formulators to sift through vast amounts of data, using this to influence their approach to flavour development.”

As personalised nutrition becomes a topic of increasing interest in the nutraceutical world, there is a strong incentive for companies to tap into this potential gold mine, details Catarina: “AI can optimise time-consuming and effort-intensive use cases throughout the value chain; a great example of this is in horizon scanning during R&D through formulation optimisation. Utilising real-time consumer data can allow organisations to access truly personalised recommendations for specific audiences based on evidence, not assumption.”

In terms of supplement formulation, predictive modelling and machine learning can rapidly analyse and augment vast amounts of data from multiple sources — including genetic information, health records, dietary habits, physical activity levels and gut microbiome characteristics. This allows for a comprehensive understanding of an individual’s unique nutritional needs and health status — informing dietary strategies for the management of specific health conditions or chronic diseases.

Once a company has developed an appropriate formulation for an ingredient or dietary supplement, the priority becomes proving that the product does what it says on the tin; this is when research comes in. Teresita believes that AI plays a crucial role in running successful clinical trials and studies. “Research may be a big investment, but it can significantly enhance the credibility of a nutraceutical product or dietary supplement ingredient. When designing a study, AI can help researchers evaluate their previous findings and estimate a trial's probability for success, while also helping them to find the best way to conduct such investigations in future endeavours.”

“It will also enable scientists to determine what kind of cohort needs to be recruited for a given study. Overall, this kind of technology can save them a lot of time and resources during the research process.” Catarina echoes this sentiment, stating: “There are millions of plant food compounds; yet we only know about a fraction of them (approximately 400,000 have been identified to date). AI technology enables companies to discover bioactive compounds much more quickly and efficiently than using a manual approach."

"Machine learning can also be used to predict undiscovered health benefits that known functional ingredients may have, while uncovering unexpected ingredient associations much more quickly and accurately compared with using traditional methods.”

“AI can offer new market opportunities for nutraceutical developers and manufacturers; it allows them to reduce the cost of testing their functional ingredients while also slashing the time to commercialisation by enhancing a company’s predictive capabilities.”

Catarina also highlights that AI can play a key role in quality control: “Through the analysis of chemical interactions, potential synergies and dosage optimisation, AI can help companies to accelerate the discovery of effective and safe nutraceutical formulations.” She adds: “This technology can also indicate ways in which businesses can improve the bioavailability — and thus the absorption — of nutrients, bringing new, effective and clean-label products to market.”

Like this story? Subscribe to Nutraceutical Business Review magazine for incisive analysis, the latest news and expert-written articles from the functional food and drink industries. For more information click here.

She also stresses that post-production quality control can specifically benefit from the technology: “With AI, you can directly enhance the quality control of nutraceutical production. By analysing data from various stages of the manufacturing process, this technology ensures that employees will be aware of when a process can be optimised, while also allowing consistency in product composition, adherence to guidelines and specifications, as well as compliance with regulatory standards.”

Although AI can’t replace every process in a company’s daily workflow, it can assist them by streamlining their efforts and enhancing the way they work. When used correctly, AI can be a highly useful tool that benefits both a company and its associates.