I was shocked to read today that IBM and Slack are partnering so that Slack can use Watson to improve their Slackbot’s cognitive capabilities. It seemed like Slack may be outsourcing the one thing that could become their most valuable strength. Except as I thought about it more, I wasn’t surprised. Slack may have been forced into this deal. I never thought I would write something like this, but Microsoft has just won a big strategic round against one of the Valley’s coolest tech startups. Here is what I think is happening.

I live pretty deeply in this world. Talla has launched three different Slack bots. But ultimately, we noticed that Slack was coming under pretty intense competition — most notably from Microsoft, but also, Google has to play here someday, right? Surely at some point Allo, or something like it, gets into Google Apps. So we made the strategic decision that messaging wouldn’t be a winner-take-all market, and possibly not even a small oligopoly. It may be very fragmented. Better to take an approach that plays to that fragmentation, which is why we ultimately made a different bet on a cross-platform ChatOps tool for HR teams.

I’m sure everyone in the market has started to realize this as well. So what I think has happened is this: IBM needs messaging and collaboration. They missed email as Lotus faded, and needed something new, and I bet they look at Slack as an acquisition target. But Slack probably won’t sell, unless the future turns against them. IBM is betting it does, and a strong partnership gives IBM a seat at the table for that discussion.

Microsoft stayed platform agnostic and wanted to focus on a playbook that made intelligent tools and their developer ecosystem strengths, while they waited to see how this would all evolve. Once the future became clear, they realized they needed a messaging player for work, and best to build Skype into that.

Slack has focused on the wrong things, believing their strength is in their hip design and early lead at integrations (and they also say Search, but I have a really hard time believing they believe that). They made a strategic mistake with their Slack app fund, investing in bot companies rather than bot creation tools. They made a good strategic move hiring Noah Weiss and their AI team, but it might not be enough. I assume that team was probably moving too slow, found themselves way behind Microsoft on AI, and probably got early wind of what Skype was doing. Add the Facebook for Work launch and suddenly you have two competitors who have distribution that is tough to match, even if you have a slightly better product.

So what do you do if you are Slack? Your best asset is your data, not your design, not your user base. I have no idea what the deal is with Watson, but I’m sure IBM sees this data set as a way to improve Watson’s NLP capabilities even further, and IBM probably negotiated some great rights around using Slack data sets. Slack’s best chance was to share their most valuable asset and hope that by outsourcing a bunch of their AI, they can fill the gap until they catch up to Microsoft and Facebook, or until they can figure out whether Skype and FB for Work are going to have an impact. But if this all moves against them, it could end up being a death knell. (OK, maybe not a death knell, but surely the coming of Slack-as-mediocre-player.)

I know a lot of people at Slack, and they are pretty smart. I don’t have first-hand knowledge of this deal, though, so my reasoning here is all speculation as an industry insider who has no real knowledge of how each party thinks about this situation. I’ve always been particularly impressed by what I’ve read about how Stewart Butterfield thinks, so I assume he has a pretty good plan. If I had to predict what happens, I’ll say this.

I think FB for Work doesn’t make it. I think Microsoft’s new version of Skype is a strong competitor. I think Butterfield goes to Google and cozies up and Slack does really deep Google Apps integrations. This gives Google a solution, and gives Slack bargaining power if IBM or Google wants to acquire them — there is a big entity on the other side of the table who can pay up. Then Slack watches to see how the chat messaging value chain ultimately develops as the industry matures. If Slack ends up well positioned and owns a valuable piece of the chain, they stay independent. If not, IBM wins and acquires Slack, and Google buys Atlassian so that Hipchat can go to Google Apps, and Confluence and Jira join the Google Cloud division. Pure speculation at this point, but still fun and interesting to think about.

This post appeared originally on Medium. Used with permission.