My latest book: Twin Wolves: Balancing risk and reward to make the most of AI
This annoying Slack notification doubles as a lesson in building AI-backed product features.
You've probably seen this pop up in a Slack channel where several messages have gone by. I feel like I'm seeing it more and more these days. And it's troublesome.
AI-based summarization tools have been hit-or-miss; so if several messages have gone by, why would I rely on a genAI bot to capture what I need to know? How does the bot know what's relevant?
Here's the thing: it doesn't.
This is especially troublesome when it appears in a DM channel. If someone has sent several messages directly to me then it seems I would want to read them all and understand what the person is saying.
What are some takeaway lessons for your company's AI-based products?
▶️ 1/ Look at the AI-based features you encounter in other products. ◀️ What do you like? What do you find intrusive? Would you want your end-users to feel the same way?
▶️ 2/ Be mindful of how you trigger notifications of new features. ◀️ Even if you think a feature is great, your end-users may not share your enthusiasm. I've seen the Slack offer appear after 5-6 messages have arrived in a channel.
(Even more insulting is when the notification arrives after I've typed out a handful of messages. Why is this app offering to summarize my own words back to me?)
▶️ 3/ Consider the feature fit. ◀️ In this case, if I've missed that many messages (say I was out of the office) I can first check other systems for decisions that were made in my absence. After that, I can rely on a team member to bring me up to speed. That person has the necessary context. The Slack bot does not.