Microsoft has recently shifted all of its focus to AI, particularly their “Copilot” strategy. When they brought OpenAI’s ChatGPT to Bing, they finally had a chance to move from a distant second in the search engine game. On the heels of this excitement, they locked up OpenAI, becoming their “Compute and Commercial Partner” with a multi-billion dollar investment. Then what happened? Let’s see.
The First Phase
Let’s face it; Bing was an “also ran” since its inception in 2009. Only able to garner less than a 3% market share vs. Google’s 90% plus. To be fair, Microsoft used Bing in more ways than just a search engine competing with Google, incorporating it into other parts of Microsoft’s technology stack. Otherwise, I suspect it would have gone the way of Nokia. For some reason, while Microsoft has been quick to dump motions where they did not feel they could win, they hung on to Bing.
Fourteen years later… a glimmer of hope. Incorporating ChatGPT into Bing sent shockwaves through Google, which was left scrambling to respond. After waking from their perpetual naps on their desks, the Bing team finally had something to be smug about. But they poked the bear. Google is not “Bob’s Search Engine and Nik Naks”; it’s freaking Google! With a 90%+ market share, and Microsoft is trying to lift a leg and pee that.
It reminds me of some advice my Dad gave me when I was a kid, “If you ever find yourself in a position where you think you may get into a fight, punch your opponent in the face first, as hard as you can, when he is not expecting it”. Well, Microsoft punched first and hard, and Google reeled… for a minute, then got to work playing catch-up quickly. But Microsoft also followed my Dad’s second part of that advice, “Once your opponent is stunned by that first punch, keep punching, don’t let him get up to fight back”.
This fight will continue for some time, but finally, after fourteen years, it is actually a fight. In the meantime, AI’s ability to bring a corpse back to life infected Microsoft and every division in it.
The Infection Spreads
While the AI motion at Microsoft is now company-wide, I focus on Business applications, so that is where I will go next. While the business applications marketplace is much more fragmented than search, there is still a market leader… Salesforce, with about a 20% market share, is far and away the undisputed King of the Hill. Microsoft’s business applications, led by Dynamics 365, have only been able to scratch out about a 3% market share. If Microsoft can shake a 90% market share leader like Google, turning the turret to place Salesforce in their crosshairs seemed like a logical next step for the AI wave.
Like Google, now Salesforce is rocked on its pedestal as Microsoft “slaps” AI on everything in sight. Almost every other project underway is moved to the back burner as AI consumes all of the oxygen. Microsoft can literally taste victory against its Bizapps Arch Enemy. But again, this is not “Gill’s CRM and Steakhouse”; it’s freaking Salesforce! Somebody at Microsoft must have known my Dad because they again threw the first punch. Adding insult to injury, they didn’t just slap AI on Microsoft’s Business Apps; they smeared it on Salesforce too.
Salesforce just announced its AI intentions this week in an attempt to shake Microsoft off its leg. But, while I said my focus is Business Applications, Microsoft has spread AI across all of their other stuff, like Microsoft 365. And what do most Salesforce users use for Productivity apps like email, etc.? And where does all of Microsoft’s AI stuff feed back into? Microsoft, via Azure OpenAI. Like the Tesla charger controversy, Salesforce’s AI can only plug into Microsoft on Microsoft’s terms. So Salesforce’s AI in Salesforce may be isolated… unable to engage with all of the other AI in the other Microsoft apps in that customer’s environment!
Brilliant or Devious?
Microsoft’s ambition to “keep punching” by very quickly incorporating AI via Copliots in everything they have, means that the first wave was… “ornamental”. Add the AI checkbox to the feature list of everything and check it ASAP. Once that is done, take another pass through and add actual user value. Do this quickly before the users realize ornaments… are just that. This will take much longer than the ornamental checkbox pass. But they have some time.
One leader recently said they felt they had a two-year head start. This was said before Google and Salesforce announced their AI initiatives, so who knows if two years is still correct, but clearly Microsoft has a head start and continues punching. It is actually one of the few times that Microsoft has had a head start in almost anything, and Satya must be giddy at the thought of not having to play catch-up. But this is also not a position that Microsoft has historically played from. Punching to stay on top is a lot different from scratching to catch up. To be fair, Microsoft has a pretty good track record in scratching to catch up, having been late to the cloud but proving, in the end, they were the tortoise.
Example of an AI “Ornament”
Microsoft launched a Copilot for Power Apps. From a natural language description, Copilot, using AI, will build a single table and create an app for it. For a person with basic skills, this will save them about 5 minutes and build a Shitty Little app. This… is an ornament. More like a Proof of Concept of a Minimum Viable Product than anything of real value to an organization. In fact, this may be the opposite, as it could accelerate morons adding to the Mountain of Shitty Little Apps even faster.
Example of AI “Value”
A Tweet led me to this blog post from Microsoft. The first sentence is, “Business leaders are asking us; What can generative AI do for my business?”. Microsoft might have edited the sentiment a little, as I think the first sentence was more like, “What can generative AI do for my business? Because so far, I ain’t seeing shit”. Regardless, Microsoft does a pretty good job in this post of outlining for a specific role, “Marketer at a Consumer Goods organization”, how some of these capabilities could assist that person. They do go off-track here and there, like what do recipes have to do with anything, but still, there are some nuggets.
Again, Microsoft’s historical position has been “catching up” with the unique advantage of “seeing” what others have done. They now find themselves in the unique position of having to “create” what others will see in a brand-new space. They’ve lost the advantage of analyzing their predecessor’s mistakes; they will now be the ones making mistakes to inform their successors. Hopefully, they have fully embraced the “fail fast” and “move past” mindset, as the stakes have never been higher.
While ornaments are still being slapped on things, previously slapped-on ornaments are being reimagined as legitimate value providers. While under NDA, I have seen some pretty impressive stuff, and I am not easy to impress. Expect to see the “Preview” tag for a while, as environments will start seeing updates almost daily. Make no mistake; it is a “race”. Microsoft was able to get a headstart using a wheelchair as they caught everyone sleeping, but the competition is now running full-sprint to catch up, and Microsoft has stood up from the wheelchair and is also running. Imagine… Microsoft… in the lead, running and punching at the same time. The mind boggles.
Microsoft later acknowledged that Bing was actually using ChatGPT 4 before anyone knew it existed beyond rumors. While Sam Altman has said ChatGPT 5 is not even being trained and won’t be for quite a while… do you believe him? Is Microsoft again working behind the scenes with the next big thing? No one else is. Will Microsoft, pouring billions into OpenAI, throw another massive punch soon? I have no idea, but I would not be surprised. That reminds me, I need to cash in my GOOG and CRM and buy some more MSFT…