SaaS is not Dead, yet

Software-as-a-Service is, according to Microsoft CEO Satya Nadella in a podcast interview last year, dead. What does this mean? Is Microsoft 365 not going to be a thing anymore? Rest assured, Microsoft 365 is going to be a thing for years to come, but I do think Satya is on to something in that the ongoing AI Agents race will eventually diminish SaaS relevancy.

The takeaway point from the podcast is that as we more and more use natural language to work with our data the SaaS backends holding our data and compute will eventually become irrelevant.

As coding agents become better and better we are seeing an increase in software being produced. This new software tends to still be of lower quality, and I suspect often should not be considered Enterprise-grade, but will probably in the short term spur an influx of more specialized SaaS offerings with appealing prices. Microsoft can easily tackle this by competitive pricing and continue to improve their AI Studio, Power Platform, Dataverse and SharePoint offerings for user self-service of business automation while maintaining Enterprise-grade security and compliance.

In the longer term, however, as open-source produced using AI agent coding support continues to flourish, the training feedback loop of AI models will eventually crystallize a preferred frontend, backend and deployment pipelines for them. Once this has been established we will get the true tests if Microsoft 365 and other SaaS offerings remain relevant.

IT in general might eventually become a commodity, making it pointless for hosting providers to offer anything else but what the AI models expect. We might simply get de-facto standards from the reinforcement learning training loops. IaaS with agents in front is thus what kills SaaS.

My bet is that Microsoft 365 will remain relevant the longest of all the comparable SaaS offerings. I have many many reasons for that bet, which I can explain if you ask me, but, in the end, 10-20 years from now (or 30, I have no idea), I do suspect much of what we currently get from Microsoft 365 will be seen as commodities.

What does this mean for company decision makers?

Before Donsense became a favourite word of mine, I used to say that if a company makes the choice to use Microsoft 365 they should go all-in on it. It does not make sense to also use competing SaaS alternatives for workloads that Microsoft 365 already cover. Now, because of donsense there are also geopolitical factors to consider. Regardless of those, it makes more sense than ever to minimize your use of SaaS that is not compatible with Microsoft 365. After all, once Agentic IaaS is all that matters you want as few technical areas as possible to migrate from.

Furthermore, continously work on your IaaS and open-source strategies. If you have none, start. Just because Kubernetes and PostgreSQL is a thing now does not mean it will be when paradigm shift happens. Figure out the balancing act of continuing to deliver business value on top of Microsoft 365 while maintaining deployment competence in line with your IaaS strategy, and with acceptable level of cannibalism where there is overlap.