I’m still getting my head around ChatGPT—a tool that somehow made Bing sexy.

We’ve seen over the past three months how training generative AI tools against a large but generic corpus of knowledge like the web produces impressive results. No argument here, but what about other, more focused bodies of knowledge?

Specifically, what about training a language model against all of the text, images, links, conversations, and videos in a Slack Workspace or an instance of Microsoft Teams?

Developers are already thinking about ways to fuse the two technologies together.

Example

Say that you just joined a company—one that eschews e-mail and uses an internal collaboration hub instead. Imagine being able to access all of the valuable content but in a more conversational, less overwhelming way. (Parsing through months-old messages to get a feel for the organization, team, department, internal politics, and major decisions is akin to drinking from a firehose.)

Here’s a generic work-related question and ChatGPT’s answer, but I can see far more specific ones that account for a team’s or a department’s interpersonal dynamics:

I could see typing a DM to Steve, only to have AI tell me that he typically doesn’t respond for three days before I send it. Maybe a bot will suggest that I search an existing Slack or Teams channel for my answer or ask Lucy or Pete instead? That recommendation might help alleviate tool overload and solve the search dilemma.

Such functionality isn’t a pipe dream. It turns out that developers are already thinking about ways to fuse the two technologies together. (A developer wrote this bot in Python.) Ditto for the C-suite. Microsoft CEO Satya Nadella has already said that ChatGPT is coming to all of his company’s products. For their part, citizen developers can integrate Slack and ChatGPT via Zapier.

Simon Says

Will generative AI improve the utility of internal collaboration hubs? Yes, but don’t ask me when. The possibilities are endless—and worth keeping an eye on.