I Went To The Microsoft AI Tour. Here’s What I Learned

I Went To The Microsoft AI Tour. Here’s What I Learned.

Last week I attended the Microsoft AI Tour at the Javits Center in New York. I officially went to a few sessions, unofficially peeked into a few others, sat in on Executive VP Scott Guthrie’s keynote, mingled, walked the floor, had a sandwich, drank three cups of coffee, hit the men’s room twice and still made my 6PM train back to Philly.

Microsoft did a great job. There were, like, a thousand people there. Probably more. Check in was fast with a QR code for those of us that pre-registered. The event was free, and it included a continental breakfast and lunch. There were plenty of smiling staff. And plenty of perky Microsoft employees, giddy with the Kool-Aid they’ve been drinking. My only disappointment with the event was just how scripted the presentations were, leaving little room for the presenters to add their own color. Oh well, sign of the times. Maybe the bots will do better when they run these conferences in a few years.

More importantly, I was there to learn about AI and how Microsoft – one of the world’s largest tech companies – was going to change my life and my business with it. I learned a lot. Here are five takeaways.

Small businesses can now develop Microsoft AI solutions.

A year ago, this was an impossibility. Not so much anymore. Microsoft has rolled out more than1,800 AI models on GitHub (yes, even DeepSeek we were told with a wink), countless other no-code/low-code apps, tools to work with more than 1,400 plug-ins on Copilot Studio and an upgrade to Fabric which brings together data from multiple places including texts and images and gives even the most inexperienced developer the ability to write queries using Copilot prompts while building a “lake house” of information. In the hands of the right developer this can be used to connect in real time and bring in and train vast hordes of data from other systems – ERP, CRM, HR, you name it, even custom databases – into a company’s own large-language-model.

Businesses with the capital and know-how can then use this stuff to build everything from a human-like customer service chatbot to an internal system that makes recommendations for inventory planning, project management, logistics, or how much drywall is needed for ten new residential homes under construction.

How much capital? A lot. Even though these tools are affordable, the overall investment wouldn’t be cheap. You’d still need a $200K per year developer, along with a few external Microsoft consultants to use them all. But this was not an option before. Now it is.

The Microsoft AI publicists are obsessed with the word “human.”

Every presenter took pains to say how all these AI tools will enhance human productivity, improve the lives of humans, create better human workers because hey…it’s all about us humans. Let’s go humans! We rock!

But only a very few of the human attendees at this event actually believes this. Why? Because when a Microsoft manager jumps into Guthrie’s keynote (it was scripted, of course) to gleefully demo how a very, very human-sounding and looking bot answers questions and helps a visitor buy camping equipment we can all read the writing on the wall. And it’s bye-bye humans! Sure, this stuff is great for consumers but we know the “humans” that actually benefit from this technology and it isn’t the customer service reps that are currently doing this work and will soon be out of a job. It’s their former employers who will be counting the money saved. When will these big tech companies finally admit the truth?

Agents are the buzzword of 2025’s Microsoft AI messaging.

Second to the word “human” was the word “agent.” They’ll be doing everything for us, behind the scenes, underneath the bed, in front of our faces…whatever. I watched demos and presentations showing how agents will soon order products from Amazon, do payroll, send out email pitches and go after malware. Microsoft is releasing a bunch of new agents for its Dynamics product lines this year, and this is just the start.

Of course no business manager in their right mind will actually rely on these agents during this first go-around because only fools rely on a version 1.0 of anything Microsoft releases. But these will get better, more accurate, and more reliable over the coming years and this event made it clear for me that that agents are the future of AI in business.

Microsoft own people still don’t trust Microsoft AI!

In two sessions – and I will not disclose which ones because I don’t want nice people to lose their cushy jobs – the audience and I watched with knowing glances as the presenters made reference to how “brave” they (or a colleague) was by actually demoing their Microsoft AI application live. What? This is brave? Glad to know I’m not the only who feels that way!

If they’re so brave doing this in a demo, how brave do their customers need to be when it’s real life data? To me, it’s still evidence that these applications – while great and exciting for the future – are still not ready for prime time. Maybe that’s good news for all those customer service reps who will soon be replaced once this technology actually works. It gives them time to find new jobs…or skills.

Finally, I can confirm that no one knows how to pronounce Azure.

It’s the elephant in the room. And it begs the question: who thought up that name? How did Microsoft’s image and brand team approve this? The Oxford Dictionary says that Azure means “bright blue in color like a cloudless sky” and I get the connection to Microsoft’s cloud platform. Bright, blue, cloudless, smooth, beautiful, etc. etc. But is it Az-ir or Az-ore? I heard it pronounced both ways by actual employees at Microsoft multiple times. Geez.

I joke, but the Microsoft AI Tour was not a joke. It was a productive day. Microsoft is putting out some serious AI stuff and if you’re running a company of any size you’re a fool to ignore it. Talk to your software vendors. Get training. Watch a few YouTube videos. Bring in a consultant. One thing’s for certain: your best competitors will.