Personalised Nutrition and Health From Consumer’s Perspective: How Whisk’s Technology Helps People Eat Better

“You are what you eat” is a saying as old as time but never seemed more true than it is today. As our understanding of how our diets can impact our health, particularly in the management of lifestyle-related illnesses such as diabetes or cardiovascular disease, changes, it is more important than ever to provide the information to help guide consumer’s choice of food.

While the science expands, consensus is hard to find with many ideas and well-researched theories competing for attention; from low fat, high fat, low carb, low calorie, high fibre, paleo, keto or Atkins, the list goes on.

Even government-approved ‘healthy eating’ definitions have come under the spotlight as many believe the science behind them was promoted by the sugar industry, is not as robust and objective as was first thought and is the underlying issue behind the obesity & diabetes epidemics.

Throw into the mix the insight that each of us can have a different response to exactly the same diet based on factors from muscle density, age, genetics or gut biomes and the idea of personalised nutrition enters the debate.

One thing is certain: anyone diagnosed with a health condition who gets told to eat “a healthy diet” can’t be blamed for being baffled and confused on how to change their eating habits.

The solution is complex but Whisk’s AI-powered technology is ready to help.

There are 5 big questions Whisk’s tech can help answer:

  • What is actually in my food?
  • How do I find a ‘healthy’ recipe for me?
  • Will I like how it tastes?
  • Why is it so hard to shop for the right ingredients?
  • How will I get personalised recommendations to help me manage my health?

1. What is actually in my food?

The Food GenomeTM is an extensive ontology that maps the world of food in great detail. To add nutritional data manually to so many nodes of data would be a lifetime’s work but Whisk’s deep-learning based natural language processing matches nutritional data to the ontology at high speed and accuracy.

Whisk’s smart, highly-visual admin suite then allows our data team to check and perfect the nutritional values which is the essential step to making accurate nutritional calculations. Getting a single value wrong of one small but important ingredient, like butter, can throw out the whole calculation for a recipe so every data point is important to get right.

Armed with that information, Whisk can then go on to calculate detailed nutritional values and Recommended Daily Amounts (RDA’s) for recipes in real time and at least as accurately as a highly-trained human nutritionist.

2. How do I find a ‘healthy’ recipe for me?

Once you know what’s in your food, the next question is how to find the right recipes to suit your preferred diet or health needs from the millions that exist on the internet or in books.

This is definitely about quality vs quantity as finding a million recipes is easy but finding the one that perfectly fits what you fancy for that precise occasion is priceless.

With the detailed understanding of the nutritional composition of millions of recipes, Whisk is able to create rules from very simple (nut free) to very complex (a “Mediterranean Diet” or Glycemic Index) that then allow us to surface recipes to your specific needs.

diatery preferences

Add to these filters the ability to search by cooking difficulty, main ingredient, cuisine, dish type, etc and very quickly you get down to a very manageable number of recipes.

The other way technology can help simplify the process of finding “healthy’ recipes is to distill complex data and numbers into clear, useful insights about food that make it easy to understand what you’re eating.

Personalised nutrition and health

Simplifying all the complex numbers and data points in the macro and micro-nutrient breakdown earlier into the Whisk Health Score is a great example of this technology.

3. Will I like how it tastes?

Now here’s where things get really interesting because, once you’ve found a recipe that looks good, it’s still hard to know if it’s really going to be a flavour you love and more importantly one your family will enjoy.

This sounds like a small issue, but it’s one of the main reasons we stick to the same 6 or 7 recipes on repeat and get stuck eating the same food – whether it’s good or bad.

One of the datasets Whisk has mapped to the Food Genome is organic flavour compound data so we have an understanding of different food flavours and how they combine and complement each other.

This opens up whole worlds of possibilities. Understanding flavour allows us to recommend flavours you may not realise you like. We can suggest foods that complement each other or ingredients that transform flavours you don’t like so much.

We can also suggest dishes that go well together, food or drinks that complement each other.

recipe and wine pairing

4. Why is it so hard to shop for the right ingredients?

Trying new recipes is great in theory, but too often when we cook on repeat, it’s because we also shop on repeat. Taking out the friction from the whole cooking process and making it easy to buy the ingredients you need is something else AI can help with.

Whisk’s deep-learning based AI parses recipe ingredients, understands exactly what they mean and then matches them to a product at a store.

As with nutrition, matching an ingredient in a recipe to a store item isn’t a one size fits all experience. Someone who chooses organic will want a different butter to a value shopper.

Whisk can take what we know about your recipe preferences or purchase decisions and apply it to your shopping list choices, getting the right product for you and speeding up the shopping list process.

5. How will I get personalised recommendations to help me manage my health?

The next stage in the Whisk journey and the evolution of AI is where we begin to use AI to combine:

  • all the data we have about food and nutrition
  • all the insights we have about individual tastes and preferences
  • robust scientific understanding of common conditions influenced by nutrition
  • data feeds from health trackers and user metrics

Combined with AI we can then create personalised, one-on-one recipe recommendations you’ll love.

With our meal planner you can then build a week’s shoppable meal plan, using your own searches and favorites or with automated recommendations.


The reason it’s hard to eat better and break out of the cooking rut we all get stuck in is because it’s hard to know a diet is right for you, to find the right new recipes, to remember to buy the right ingredients, to have them in when you need them and, worst of all, to have a dish rejected by people you love after all that effort.

Yet nothing could be more worthwhile. Variety is the key to eating well and getting the right nutrition is not easy. Technology is perfectly placed to make it a pleasure to find new food and to add some of that joy into the cooking process.

To get your recipe content up to scratch and automate tedious data work, get in touch with us!