Project

Call for Partners | tAIstify: Leveraging AI to create tasty, healthy & sustainable foods

Perceived flavour is a critical factor in consumer acceptance of processed foods: taste remains king! Food (re)formulation efforts aiming for healthier and more sustainable products are often accompanied by undesirable flavours. Understanding of, and arriving at optimal flavour profiles is time-consuming and expensive. Artificial intelligence (AI) with appropriate data and knowledge, can greatly reduce this effort: it streamlines the formulation process, mitigates risks, and therefore expedites market entry, boosts success rates, and ultimately contributes to healthier and more sustainable food choices.

The tAIstify approach

Current consumption patterns significantly affect both planetary and human health. Despite increased awareness, behavioural changes remain limited. Urgent action is required, including more frequent substitution of animal proteins and reductions in salt, sugar, and fat content in food products. However, product (re)formulation can result in unexpected and unpleasant flavour perception, lowering consumer acceptance of these products.

Flavour, a combination of taste and aroma, is one of the main drivers of food products' acceptance and rebuy. To better understand consumer acceptance and find solutions for the challenges ahead, integration of product properties, flavour analytics, and sensory data combined with expertise in reformulation, ingredient interaction, and fermentation is needed.

Artificial intelligence (AI) is a rapidly developing and disrupting industries world-wide. This project explores how AI technologies can best be applied for supporting food design and flavour optimization, to support the development of healthier, sustainable foods, accepted by consumers. We expect AI can play a role in gathering/combining novel and existing flavour information (e.g., from literature, existing databases, measurements, data from partners), and in identifying/modelling the complex relationships between product composition, sensory properties, flavour profiles and consumer feedback. A predictive model can support food (re)formulation processes to optimize flavour, and help balance the trade-offs between flavour, healthiness, sustainability and consumer acceptance. Depending on available data, the approach would be applicable across various product categories.

Use cases

Proposed directions for use cases where (re)formulation and processing may challenge product liking and acceptance:

  1. Optimize aroma and/or taste perception resulting from off-flavour compounds or alterations in flavour profiles
  2. Mitigate bitterness and balance sweetness profiles
  3. Maintain savoury and salt perception
  4. …

The use cases and product categories are selected based on the interests of the industrial project partners.

Partners

We are looking for Dutch and internationalpartners to apply together for the TKI-A&F PPS Agri & Food grant call (deadline 1 Sept 2024): Food manufacturers, ingredient suppliers, flavour houses, retail organisations