💬 AI in FoodTech

You may be part of the more than 100 million people that have already toyed with ChatGPT. I didn’t give much thought to it as I didn’t see many applications. However, I received many questions from agribusinesses and food companies. It seems ChatGPT has become a symbol for all the potential productivity increases and many concerns created by Artificial Intelligence (AI).
The number of use cases for AI all along the food supply chain is impressive, notably those. Let’s review them, with a focus on those involving ChatGPT.

Upstream, many applications in farming, notably in the large area of precision farming.
There, AI can be used in the following ways:

  • Crop monitoring, notably to analyze satellite/drone imagery and sensor data to monitor crop health and predict yields (eg. Gamaya). This can notably be then used for:
    • Sales prediction: helping farmers sell their output at the best moment, depending on how markets evolve.
    • Input management: to increase or decrease the amount of water or fertilizer (and which to use) needed by crops.
  • Livestock monitoring to monitor the health and behaviour of animals, notably to identify and prevent diseases (e.g. Connecterra).
  • Farm automation: AI can be used to pilot robots in the fields (e.g Farmwise), mostly to harvest crops and control weeds.

We have a great variety of tools in these three areas, all with their own interface. That’s where the idea of conversational AI comes to play and potentially adds a lot of value: we can imagine a situation where the farmer is managing all this information through a “conversation” and where only the most salient elements are presented to him in an easy and actionable way.

Additionally, AI is already used to create “new crops”, either through genetic engineering or traditional breeding techniques. Here, it can help to “guess” which crossing will have the desired traits.

Midstream, AI is already used for transformation (Foodscience) for many applications, such as recipe optimization. It can gather consumer preference data and create new recipes. This is yet a very young space, but it is quite promising, notably in the industry, where it’s always hard to balance cost, processing and taste.

One of the areas where AI is already used is in the growing space of alternative proteins, notably in:

  • Identifying interesting properties in nature: many plants have not yet been “explored”. Some may have interesting properties we would like to use to create cleaner labels. One example is The Live Green Co, which looks for alternatives to methylcellulose for plant-based meats.
  • Synthetic biology, and more precisely in precision fermentation: it can take years to recreate the desired protein from a bacteria and then scale the process. Multiple companies are designing faster processes to go from the idea to semi-industrial scale. Eventually, at scale, this would be really game-changing, enabling any food company to identify a protein with a desired property (let’s say a binding agent or an egg protein) and “order it” to then experiment with it in a few weeks.
  • Managing bioreactors, notably for cellular agriculture. My companies in this space have demo products, but none have yet the ability to scale their production. Startups are trying to solve this issue by creating smart bioreactors to once again speed up this process.

Downstream, and closer to the consumer, the use cases for AI are once again abundant. Most are invisible to the consumer, such as:

  • Quality and food safety controls
  • Data sharing (a great pain) between suppliers and retailers
  • B2B marketplaces (one of the hottest topics of this year)
  • Supply chain optimization
  • Food waste, both in retail stores (to create some kind of yield management to adjust prices in real-time when fresh products are going to be wasted, such as Smartway) and in restaurants (through image recognition to detect what is thrown away and then to better order)

However, in some instances, AI can be used in consumer-facing applications:

  • For transparency: to bring the most relevant data to the consumer about the food items he buys (or should buy).
  • Personalization: to analyse consumers’ health and dietary preferences and then create personalized nutrition plans.

Here, in both instances, the use of conversational tools could greatly improve the efficiency and even the observance of the recommendations by the user.

Many other use cases exist. In a word, AI is everywhere and nowhere at the same time. This technology is only a mean that all players can use to have a greater impact, reduce cost and/or improve the efficiency of the solution they are working on. To succeed, in all instances where technology is facing its users, it has to be invisible. That’s where a conversational tool (through text and soon through audio) such as Chat GPT makes a big change.

My bet is that the impact of AI in food will be massive but quite slow to materialize as it requires adaptation and often coordinated change all along the supply chain. However, established players have considerable opportunities to differentiate themselves right now. A first step would be at least to list all the areas where their current business could be disrupted or improved through AI, now and in the future.

Have a great week,
Matthieu

13 notable deals and things to know this week (2023 - week #9)
8 notable deals and things to know this week (2023 - week #8)

Leave a Reply