This week marks the passing of Watson Workspace, one of the few major innovations in the ICS space of the last decade. I say that because it was innovative not only in technology, but openness, open technologies adopted and showcased, design thinking approach, integration and most importantly user experience. It may be gone for good, but for me it did a lot of good.
There are a lot of question marks still left around the decisions of the last six months. That’s not least because communication from ICS has been a paradigm in worst practices and the complete antithesis of what we’ve seen for first Connections Pink and then Domino. Indeed even though it was an open secret, an elephant in the ICS community room, IBM Champions were forbidden from speaking about it to customers for months. Our very vocal criticisms of that policy went unheeded, our attempts to understand were ignored. ICS, as it transitions to HCL Digital Solutions, will not be tainted by this debacle and I have no qualms about getting involved in innovations that fit with the core direction I see HCL going; IBM will have to work very hard if they wish to remain within some of those customers.
The following “why” is fully speculation. Shortly after the products were withdrawn from sale, the HCL acquisition was announced. All other ICS products were moving to HCL, question marks were raised around Kenexa products which are now moving to the Cloud parts of IBM, leaving Watson Workspace as a unique element. Let’s be clear, regardless of the proprietary “Watson Work Services” API calls, there are massive numbers of calls to Watson services of various kinds (sentiment analysis, Watson Assistant, potentially translation and surely more). So no company – HCL, business partner, whatever – was ever going to buy Watson Work Services. The other Watson “products” like Watson Assistant and Watson Translation are APIs only, for use by developers in other products. And the products themselves like Watson for Oncology and Watson Discovery are not aimed at generic end users. IBM seems to be getting out of that space and focusing on specialist big data products and cloud services only. So regardless of how well Watson Workspace may have been, it’s possible IBM would have pulled the plug regardless. Watson Workspace doesn’t fit into the direction IBM seem to be going. Is their strategy right? I don’t know, but this week Volker Weber talked about another big change of strategy from Big Blue, when Lou Gerstner bought Lotus in 1995 and brought IBM back from the brink. That got IBM very visible with all customers, gave them software in a lot of customers. This strategy will ostensibly not, even though IBM has a presence in much open source software that many customers have.
Why Wasn’t Watson Workspace More Pervasive?
When Watson Workspace was announced at World of Watson 2016, it made a big splash. It didn’t have everything that the competition like Slack had, but it was good enough to catch on with this community and developers. When Inhi Cho Suh took over from Jeff Schick at IBM ConnectED 2017, she continued the focus on Watson Workspace and I think it’s telling that Anne-Marie Darrough became the Offering Manager for it. IBM Think 2018 promised a new dawn for Watson Workspace. By then Microsoft Teams had moved much quicker to release a product that got much more marketing, so often the Achilles heel of IBM. But the announcement of the Zoom partnership made a big splash, rightly so. I can’t speak for the Zoom integration from a sales point of view for Watson Workspace Plus, but as a consumer it was a great move. The other announcement was templates. Again, this was very promising, with the right enablement tooling. Key to that was allowing business partners to demonstrate to customers the effectiveness, but that also required business partners being able to understand what made a good use case and the effort involved. There were a lot of synergies with chatbots, with integration of many services and in most cases needing some kind of application in between. Domino was an obvious choice for that, with the Watson Work Services Java SDK I had been involved in or the Node-RED nodes Stefano Pogliani had done such a great job with. Unfortunately, we never got that far. The off-the-shelf templates never got released, templates could only be created or modified via APIs, so they never became customer ready, in my opinion. But the chat side had also matured significantly. That was at the time the product numbers were pulled, shortly after IBM and others had delivered sessions at Social Connections last year, sessions that could have been offered to other products, at a conference that IBM pulled financial support for late on.
Not All Bad News
But IBM did get a lot of things right with Watson Workspace.
The user experience was a big win. When alternatives were investigated, Skype was very unpopular with many people. I’ve seen a lot of less than positive comments for people having to move to Slack. Zoom has a chat and makes a lot of sense for the video aspects. It’s chat is behind the competitors, but it’s an aspect Zoom are looking to improve and already have in the last few months.
The API-driven approach was a very big win. Being able to programmatically do everything that could be done in the UI brought a lot of power for developers, as we saw with the Watson Work Services XPages SDK and the Watson Workspace Client for Eclipse. It also allowed IBM to build an export tool, without which conversations would have been lost completely. The export format is far from perfect, but at least the content is available.
The use of GraphQL was a great innovation. I’m glad it introduced me to GraphQL, which I’ve spoken on at various conferences. GraphQL makes a lot of sense for Connections, and this project interests me greatly. It’s a great fit for NoSQL databases and I’m sure we’ll see it at some time for Domino, I know it’s come up at various jams. There are some quirks of Domino over something like Connections, but I can envisage approaches for that. And GraphQL was announced this week for Portal Digital Experience. GraphQL, quite frankly, is a no-brainer if you want to produce a REST API that’s flexible.
The reason templates interested me is that integration of chat and applications brings a lot, especially when those chats are available in context alongside the data. I’m sure we’ll see it come back in the future.
Sametime has persistent chats. Obviously nothing can be said at the moment. So we’ll have to see whether HCL want to keep Sametime a good product only for Digital Solutions customers or whether there’s any intention to take it wider in the way Watson Workspace did, with a free edition. But regardless of the future plans, something is clear: Sametime was there before Watson Workspace and will still be there after Watson Workspace.