Improves customer experience with users adjusting recipes 15% less because AI can predict what they have in their cupboards
Increases customer engagement and reduces failed searches through Vertex AI Retail Search
Offers the opportunity, in future versions, to deliver data insights across the organization reducing demand on analyst resources
Cherrypick encourages healthy eating and reduces waste for budget-conscious shoppers with AI-powered meal planning and shopping app.
Cherrypick encourages healthy eating and reduces waste for budget-conscious shoppers with AI-powered meal planning and shopping app.
AI: A helping hand in the kitchen and the supermarket
In the world of online grocery shopping, the time-consuming and overwhelming process of selecting dozens of items can often lead to poor dietary choices. Using Google Cloud, Cherrypick is changing that for over half a million customers by making healthy food simple with thousands of AI-personalized recipes.
The company recognized that convenience food laden with ultra-processed, calorie-dense ingredients was creating unsustainable patterns in both personal health and environmental impact. Yet existing meal planning solutions lacked the sophisticated personalization needed to change consumer behavior, resulting in disengagement and missed revenue opportunities.
Chris Parsons, CTO and Co-founder of Cherrypick, says, “We realized that convenience food was shaping poor health outcomes, but cooking at home was just too hard for many people. We wanted to build an AI-driven experience that made healthy choices as easy as unhealthy ones.”
From a technical perspective, Cherrypick wanted to scale the AI capabilities of its app without compromising profit margins. Early experiments with a WhatsApp-based grocery assistant demonstrated this challenge vividly. While technically feasible, each customer visit involved multiple AI queries, making the service too costly to be profitable. The team needed a solution that delivered personalization while remaining economically viable.
Google Cloud enables us to deliver AI features at scale without sacrificing speed or cost.
Chris Parsons
CTO and Co-founder, Cherrypick
Building a healthy community fast
Cherrypick's transformation began with a carefully engineered AI solution built on the Google Cloud technology stack. At the core of their implementation, the team deployed a customized machine learning model that estimates what customers have in their cupboards and suggests just the right ingredients for their recipes.
The Cherrypick architecture uses BigQuery for rapid data analysis, enabling the system to process customer’s purchases and user behavior patterns at scale. This analytical foundation supports the personalization engine, and Cloud Storage provides secure management and data transport layer for the completed model.
“Google Cloud's scalable, cost-effective solutions were the key to our success, allowing us to build an AI-powered experience that makes healthy eating accessible without compromising our business model,” says Parsons.
Beyond this, Cherrypick has implemented Vertex AI Search for Retail, extending it with custom metadata to support dynamic and powerful recipe searching across ingredient types. “This feature has brought the power of Google’s search to our recipe platform, enabling customers to find the exact recipe they want,” says Parsons.
Google Cloud's scalable, cost-effective solutions were the key to our success, allowing us to build an AI-powered experience that makes healthy eating accessible without compromising our business model.
Chris Parsons
CTO and Co-founder, Cherrypick
Serving up success time after time
Throughout its AI implementation journey, Cherrypick uncovered several valuable insights that have broader applications for companies deploying intelligent systems. Perhaps most significant was how to surface the model's explainability data. “We had a number of options for showing the customer the reasons why we think they already have an ingredient,” says Parsons. “In the end, we opted for simplicity. We remind them when they last bought the item so they can make their own decision.”
The team also retained the ability to override the model and tick the box of a given ingredient. “We use this to help customers get what they need and track accuracy,” says Parsons. “This gives the user a way to correct the model, and tells us how well we are doing.” With this information, Cherrypick can continually improve and train the model.
Parsons says, “In partnership with Google Cloud, we've developed an AI-driven platform that supports a healthier diet and better health, proving that technological innovation can be a powerful force for personal and environmental well-being.”
“Building with Google Cloud has taught us that technical rigor and empathy for the user must go hand in hand," says Parsons.
In partnership with Google Cloud, we've developed an AI-driven platform that supports a healthier diet and better health, proving that technological innovation can be a powerful force for personal and environmental well-being.
Chris Parsons
CTO and Co-founder, Cherrypick
“Our biggest wins came from listening to customers and iterating quickly on both product and process. The results speak for themselves. 58% of Cherrypick users now cook more from scratch, and 82% report eating better overall.”
They also enjoy a healthier diet. With 70% eating more plants, 90% trying new foods and cuisines, and 26% eating fewer ultra-processed foods. It all adds up to a better, healthier lifestyle, powered by the latest advances in Google Cloud technology.
Cherrypick is an app that helps you find, plan, and order recipes from your favorite supermarket.
Industry: Retail
Location: UK
Products: BigQuery, BigQuery ML, Vertex AI Search for Retail, Cloud Storage
