Chick-fil-A's ML Capacity Boost

Shaan shares a story about how Chick-fil-A is using advanced technology to handle customer demand far beyond their stores' designed capacity.

"I read this blog post from the Chick-fil-A tech team. They put their stake in the ground saying they're different - they're only open 6 days a week but do more sales than people open 7 days.

Their stores were originally designed to handle about 100 customers a day, but the demand is so high that they get 3 times more output than the store was designed for. They tried to predict demand better, but there were too many variables - soccer games ending nearby, weather changes, traffic on the highway.

So instead, they put hundreds of sensors inside their stores. In real time, they're taking the order as someone types it into the machine, matching it with sensors that monitor things like the fry machine activity, and giving signals to workers about what to do next. They're taking all these signals - how many people are in the drive-through and other metrics - to make predictions about what they need to make. This helps them get that extra 20% of orders done per hour for throughput.

They have about 100 smart devices in each store feeding into a machine learning algorithm, allowing their frontline workers to be more effective and get more sales per square foot."