At Shopify, we have a lot of information locked away in archived recorded presentations. There’s a lot of great information in them, but the amount of time to watch them all would mean no one would get any actual work done. And unfortunately not much metadata exists to make searching through them easy. I wanted to figure out a way to extract the goodness from inside these presentations and get them to Shopifolk new and old in an easy way.
Like many tech companies, we use Slack to communicate with one another. We’ve been playing around with Slackbots and I realized it’d be perfect if one of them could answer these stupid questions. Instead of making Shopifolk hunt for information, we could have the Slackbot bring it to them. The important part is that humans still play a role in determining keywords and potential queries, although as we collect more information we’ll potentially start applying machine learning as well.
This presentation will share the experience of bringing the Slackbot to life, including, what went right and what went wrong, lessons learned that other companies could apply, and how we plan on improving it. Making sure all employees are aligned and up-to-date is key for many companies, and I hope the work we did will help others in a similar boat.