Using AI chatbots to power a digital learning jam

At the end of 2017 I was involved with one of the most interesting learning projects I have ever designed and run.

As part of Hyper Island in Singapore we were working for Axiata, a telco client based in Malaysia but with businesses across South East Asia.  The challenge was to engage up to 12,000 staff across six countries in a digital learning event, a Digital Jam, that would be support a bigger digital transformation process that is well underway across the group.

Jams are company wide online conversations around innovation and change and were popularised by IBM in ten years ago. Hyper Island had participated in a week long DBS Jam in 2016 and proposed the format to Axiata as a way to involve as many parts of the business as possible.

Our design for this event was divided into three parts: preparation, getting ready and the Jam Day itself. Preparation involved identifying champions in each of the participating companies, designing a series of activities in which participants could engage, planning and executing a company wide communication plan and selecting a platform to support the Jam.

Getting 12,000 people to participate simultaneously on a single day was always going to be hard and so we planned a digital on-ramp to the Jam Day itself by designing a  10 day “Get Ready” process.  This consisted of a series of digital learning activities that mirrored what we planned for the day itself.  Almost all of the activities followed Hyper Island’s “learning by doing/making” philosophy including exploring VR, building a website, prototyping a mobile app and exploring data visualisations.

Digital Jam Day, the 23rd of November, was designed as a festival of digital with each operating company offering their own off-line activities and meet-ups while we planned a central “Mission”; work in teams to design and prototype the near future for the business. Once again the focus was on making and prototyping rather than the typical quiz or watching video formats favoured by many large company training programmes.

At the heart of our Jam Day we, together with Nick Oates from Other Media, designed and developed a chatbot avatar called DJ who distributed activities, supported groups and maintained scores. Our activities were gamified to encourage participation.DJ,  Our chat bot was built in node.js to run inside Facebook’s Workplace community and used the machine learning engine to detect the intent inside questions raised and to get better at supporting participants over the 10 days of the Jam. He was also starred in 3D in our printed and online awareness campaign.

There are so many learnings from this event that they deserve additional blog posts but the biggest is the power to use bot technology to support learning.  Our participants really enjoyed interacting with DJ and it was often hard, although fun, to keep up with the range of questions they expected him to be able to answer.

DJ sent out over 1 million messages over the course of our event with over 1000 teams formed and, 8700 staff members involved and over 600 final submissions.

We’re certainly looking out for more opportunities to do this again building on the power of the chat format to enhance learning.

Leave a Reply