For a nutrition platform that recommended dietary plans to customers

The Problem

One of our clients - a nutritional platform - faced a challenge that some of the meal plans that they were recommending were being rejected by customers. This was impacting sales and hurting their brand reputation. They were interested in addressing this by using an AI model that could predict if a customer was likely to reject a meal plan, and recommend alternatives in case they were.

Our Solution

We developed an ML-based quality control system within the platform's meal plan recommendation engine. The system leverages customer data, preferences, dietary requirements, and platform usage history to predict acceptance or rejection of proposed meal plans.

Data Analysis: We analysed customer demographics, preferences (likes and dislikes), dietary targets, and historical interactions with meal plans on the platform.

ML Algorithm Implementation: Utilising machine learning algorithms, we created a predictive model that assesses the likelihood of customer acceptance for a given meal plan. The model factors in individual preferences, dietary restrictions, and past engagement with meal plans.

Quality Control Mechanism: If the ML system predicts the potential rejection of a meal plan, it triggers an alternative plan recommendation. This iterative process ensures that customers receive meal plans aligned with their preferences and increases the likelihood of acceptance.

Collaborative Filtering: The system incorporates collaborative filtering techniques, considering similar customers' behaviours and preferences to improve recommendation accuracy.

Deployment: The ML-based quality control system was deployed on AWS EC2 servers.

Results

The ML-based quality control system correctly flagged 80% of recommended meal plans that would have been rejected by customers, leading to higher acceptance rates and improved customer experience.

Sample meal plan