Google Cloud, Spoon Guru to Offer AI-Powered Food Insights

Data processing and modelling is core to the Spoon Guru app, with the company processing more than 14 billion data points a day with Google Cloud. 

Google Cloud is working with smart shopping platform, Spoon Guru, to deliver AI-powered insights to people with allergies and food intolerances. 

Since teaming with Google Cloud, Spoon Guru has grown from a startup to a partner of major food retailers, on track to achieving a 75% YoY revenue growth. 

Using Vertex AI, BigQuery and Cloud Run, the company is now able to process 20,000 products per minute, tagging supermarket products with hundreds of dietary attributes. 

Within a year of its launch in 2015, Spoon Guru had signed a deal with Tesco. However, it soon found it hard to scale up. “We managed to get some significant traction, but we couldn’t scale either technically or commercially on our own,” says Markus Stripf, CPMO and Cofounder. 

“We are a relatively small company, and sales cycles are very long. For us to scale commercially, we needed to be aligned with a strategic partner to go to market more quickly. At the same time, technology moves at such a pace, we wanted to be sure we could keep up.”

“We felt our mission fully aligned with Google’s. Google organises the world’s information and makes it universally accessible – and that’s precisely what we are doing with food,” Stripf adds.

“With Google Cloud pushing the envelope with AI, there were so many potential synergies. We jumped at the chance to team up with Google Cloud.” 

Spoon Guru migrated to Google Cloud in 2023 with the support of Google Cloud Partner Appsbroker | CTS. 

The company’s serverless architecture, based on Cloud Run, now gives it more flexibility in development, as well as the ability to scale its resources to meet growing customer demand.  

“Previously there was a lot of infrastructure management, but with a serverless architecture, that’s all gone,” says Anne Marie Muscat, Engineering and Data Science Manager at Spoon Guru. 

“By giving our developers more autonomy to test new features more quickly, Cloud Run makes the whole development flow easier. And because our infrastructure automatically scales up and down on demand, we are no longer paying for resources we don’t need.” 

This auto-scaling infrastructure has allowed Spoon Guru to move away from batch processing to end-to-end processing, enabling it to process retailers’ ever-changing product data in real time. 

Since migrating from its previous cloud provider, Spoon Guru can process products four times faster, increasing its capacity to 20,000 products per minute. This allows retailers to provide shoppers with the most accurate and up-to-date product information, essential for shoppers whose allergies leave them no margin for error. 

Data processing and modelling is core to the Spoon Guru app, with the company processing more than 14 billion data points a day with Google Cloud. 

The data is stored in Google Firestore, Cloud Storage, and BigQuery, with Spoon Guru using Vertex AI to build, train, test, and deploy machine learning (ML) models to clean up, normalise, and model that data.  

“We use our data to tag products with hundreds of dietary attributes, which retailers can use to allow shoppers to search for a specific product, filter products by attribute, or find the right ingredients for a specific recipe,” Stripf says. 

“When shoppers conduct a dietary search, they spend 10% more on average. When they use dietary filters, they add 13% more products to cart. Consumer trust goes up, loyalty goes up, and it has a direct impact on retailers’ bottom line.” 

Spoon Guru also uses ML models to power product recommendations, substitutes, or healthier alternatives based on a shopper’s dietary needs. 

With Vertex AI, Spoon Guru is able to build and run its ML models in one integrated platform, giving it more agility in its model development and halving its development setup time from two hours to one. 

“Vertex AI helps us manage the whole ML lifecycle,” says Muscat. “We can build and test a new model quickly. And we don’t need a DevOps person to deploy it, we can simply deploy it through Vertex AI for production. Vertex AI makes the whole workflow a lot easier.”

Most recently, Spoon Guru has used Vertex AI to develop a genAI feature for retailers that allows shoppers to enter a natural language prompt and receive specific product or recipe suggestions.  

After migration, Spoon Guru launched on Google Cloud Marketplace, expanding its market reach by gaining access to all Google Cloud customers, as well as aligning on go-to-market strategies and sharing the stage at high-profile industry events. 

“Google Cloud has opened doors for us with their customers. It’s been a fantastic collaboration, with great synergies between our products allowing us to take a very powerful joint proposition to market,” says Stripf.

Spoon Guru has thus far won two new contracts as a result, agreeing terms in just three months, as opposed to the two-year sales cycles the company was used to. These reduced sales cycles have in turn allowed its small team to invest its time in further developing and growing the business.  

It is now looking at how it can apply its technology to new verticals, such as health and fitness and travel, with new deals said to be in the pipeline. 

“Previously we were just one of thousands of startups, a tiny company out on our own,” Stripf comments.

“Working with Google Cloud has brought commercial opportunities to grow and scale. And on the technical side, we have access to a phenomenal resource. Operating on the bleeding edge of technology is incredibly exciting.”