AI in food and beverage industry R&D: Overview
- Around 90% of food and beverage R&D leaders routinely use AI
- AI reduces development times, and costs and boosts innovation success rates
- Autonomous experimentation could account for nearly 20% of NPD time by 2030
- Companies protect formulations, trade secrets and data from AI models carefully
- Human expertise remains crucial despite faster, cheaper AI tools
AI is increasingly being used in R&D, with the majority of food and beverage companies using the technology.
While it can streamline food R&D, making the process more efficient and often cutting costs, it also comes with risks.
Why AI is being used for R&D
Food and beverage is embracing AI. According to a report from agritech consultancy Bright Green Partners, around 90% of R&D leaders in the sector are “routinely” using it.
AI can help companies streamline the R&D process. According to the report, companies such as Nestlé have shown that cutting NPD time by weeks is possible with AI. Furthermore, AI R&D is offering higher success rates and cutting costs, the report says.
Use of AI in R&D is growing rapidly. Bright Green Partners predicts that nearly 20% of the time spent on NPD could be accounted for by AI and autonomous experimentation by 2030.
Competition is pressuring companies to use AI more. According to the report, nearly three-quarters of CPG R&D leaders have seen rivals reach the shelf first, and many fear that they will miss out on this speed and capability.
Furthermore, AI is getting cheaper – in fact, costs are falling around 13 times per year, according to the report.
The caveat is that these tools are available to everyone. Humans, and human expertise, can still provide an advantage for businesses.
The problems with using AI for R&D
However, using AI for R&D is not without its problems. For example, companies are often reluctant to connect sensitive test results, formulations and IP to powerful external AI models.
Companies are therefore exercising caution. “Companies are increasingly separating sensitive R&D data from external-facing AI tools, keeping proprietary formulations, test results and trade secrets within secure, enterprise-grade environments,” says Ewa Granosik, CEO and managing director at Bright Green Partners.
Some companies are developing internal AI solutions, she says, or using specialised platforms with controlled data access as well as contractual safeguards or restrictions on using customer data in model training.
“The key is balancing the benefits of external AI capabilities with strong data governance and IP protection.”
Another problem with using AI R&D is that internal knowledge is often fragmented or difficult for it to make use of. Companies need to integrate it into existing workflows without creating disconnected pilots.
When redesigning workflows around AI, Granosik explains, it is important to establish cross-functional ownership between R&D, IT and business teams. Employees need to be trained to incorporate AI into everyday decisions, rather than simply treating it as a standalone tool.
The most successful companies start small, but design for scale, Granosik explains. They identify bottlenecks, define measurable KPIs and scale solutions only after value is demonstrated.
Need for speed
AI can radically streamline R&D. But companies need to adapt to fully take advantage of this technology.
Workflows need to be reimagined, data protected and the future planned for. Importantly, human input remains vital, differentiating a company from a competitor that is using the same AI tool.
Yet despite these caveats, AI is truly transforming R&D, and this shows no sign of slowing down.
