KFC
KFC Python django, AWS ECS, Postgres, Docker Unlike its competitors, KFC does not yet have an AI-based product recommendation system. The AI recommendation algorithm, initially deployed in 5 restaurants in Albania, was subsequently expanded nationwide. The objective is to optimize the AOV (average order value). Fenxi is implementing several initiatives: development of the product recommendation algorithm; an […]
Context
Unlike its competitors, KFC does not yet have an AI-based product recommendation system. The AI recommendation algorithm, initially deployed in 5 restaurants in Albania, was subsequently expanded nationwide.
The problem
- No recommendation engine on the kiosks, while competitors already had one.
- Average basket driven entirely by what the customer thought to order alone.
- No detailed performance tracking, by restaurant or by product.
- A product catalogue to maintain with no dedicated interface.
Solution
- Built a product recommendation algorithm wired into the order kiosks.
- Built a back office to manage the catalogue and the recommendation rules.
- Built performance dashboards, restaurant by restaurant.
- Set up containerised infrastructure on AWS ECS to hold up at peak hours.
- Ran a pilot in 5 restaurants before opening it to the rest of the country.
Our approach
Framing
Understanding the kiosk ordering journey, the available data, and what the team actually cares about: average order value.
Model
Building and testing the recommendation algorithm against order history, before anything went into a restaurant.
Build
Catalogue back office, tracking dashboards, deployment on AWS ECS with PostgreSQL.
Pilot then rollout
Going live in 5 restaurants, measuring, then extending to the whole country.
Results
The engine went from a 5-restaurant pilot to the full Albanian estate. The team tracks performance from its own dashboards and updates the catalogue without us.
Start with Compass
In 2 weeks, you know what your data can actually do for you. A structured audit, a prioritised roadmap, before you spend anything else.
A similar project?
We take 30 minutes to understand what you need. You leave with a clear opinion, even if we end up not working together.
Let’s work together
Share your need, we will get back with you quickly.