About the project.
To amplify the No Turning Back local campaign, Nike Running held a night run that was part amazing race, part escape room. The Sydney course consisted of eight secret city checkpoints, each with three clues of varying difficulty.
Our challenge was to find a seamless and innovative way for runners to trigger and receive clues – something that would elevate the event and give them an epic running experience.
What we did.
Using artificial intelligence and natural language processing, we built an SMS chatbot that could understand and respond to natural English phrases. A world first for Nike, 50 teams of four women interacted with the chatbot to unlock clues, navigate the race and find their way to the finish line.
To make the race experience even smoother, we built a separate app that used location services to determine when teams arrived at each checkpoint and automatically trigger their first clue for the next location.
As well as allowing runners to ask the chatbot how far away they were from the next checkpoint, the location services app enabled us to track each team’s location and send a hint if they were heading the wrong way.
But we didn’t stop there. The chatbot could also tell them how to find the nearest toilet, where to get water, what to do if they were hurt and what to do if they felt tired or demotivated. It even had responses to off-topic questions – and wasn’t afraid to come back with a bit of sass.
In addition to building the chatbot and app, we wrote the clues that were delivered by athletes including Gen LaCaze and Madi Robinson, and created visual clues that were sent by NRCbot. We also developed a real-time leaderboard that displayed teams’ race times, inclusive of the time penalties incurred by requesting extra clues.
How we did it.
The chatbot was powered by Amazon Lex, providing deep learning functionalities and natural language processing. We trained the bot to understand over 100 phrases that participants might use throughout the race.
We used Twilio’s Programmable SMS service to send and receive text messages.
We built an iPhone app in React Native. The application was sideloaded onto 50 phones where it resided in the background, notifying a central database of location changes every two seconds.
We created a dashboard in React to orchestrate the event entirely using serverless technologies. Various AWS services were combined to achieve this, including AWS Lambda, AWS API Gateway, and Amazon DynamoDB. The dashboard featured a live map, which allowed us to track the teams in real time and automatically trigger message sends at key locations.
The race journey.
The teams arrived at Sydney’s Carriageworks with no idea of what they were about to take on. After picking up their race kits, they headed to the Get Connected bar – where all was revealed. Each team was given a personalised N+RC phone and invited to start interacting with NRCbot.
A video message from a Nike marquee athlete got teams ready to run and the first clue was released – signalled by a chorus of message tones.
The teams had to solve the first checkpoint clue in order to escape Carriageworks, before heading out into the night with only NRCbot’s messages to guide them.
Their reward for embracing the unknown and reaching the finish line? An epic celebration at the MCA’s stunning Sculpture Terrace.
1,000+ messages sent
80 clues delivered