EP 6: Microsoft Build 2023: Top 5 Teams and AI Announcements

Transcript for this Episode:

INTRODUCTION VOICEOVER:
This is Tech UNMUTED. The podcast of modern collaboration – where we tell the stories of how collaboration tools enable businesses to be more efficient and connected. With your hosts, George Schoenstein and Santi Cuellar. Welcome to Tech UNMUTED.

GEORGE:
Welcome to the latest episode of Tech UNMUTED. We're going to take a little different approach today. We're going to drill down on a couple of the key highlights that came out of the Microsoft Build conference today. Specifically, there was a keynote, and then there was a follow-on conversation that took place around AI and some of the new tools that are coming out and we're going to break some of those down. The focus is more from a business productivity standpoint then some of the deep dive technical stuff that was in there, we might hit on some of that but it goes a little deeper than we're normally going to go on these podcasts.

SANTI:
Yeah, and that's because Microsoft Build, the build conference is really geared towards the developer community, right?

GEORGE:
So I want to share a couple of slides and guide the conversation. If you're listening on Spotify or another podcast platform, you'll be able to follow along just as easily. We're going to narrate around all these slides as we go through it. This is just an image early on in the conference. This was a live conference Santi and I happened to watch it on video. We weren't there, but there's a couple of things they kicked off with, one that was sort of interesting was effectively the productivity impact of different technologies over time. This is a scale that goes way back, right, it's got the printing press and then it goes to a bunch of other elements revolution wise from a technology standpoint. But that last bucket of stuff is really where AI is and some other things, and we're going to hit on another slide in a second to really break down the key stuff that Santi and I, you know, personal experience and what we felt happened in there. The other piece on the bottom is – there were incremental improvements in time from – it's technically positioned as GDP but these are effectively productivity increases, so things overtime got more and more productive. And when you look at some of those inflection points, we're going to hit on each of these quickly just to give again our experience but one of the key things is the speed of change has accelerated, so things that took years and years, right, even decades, to really get fully ingrained, are now happening in a matter of months when you look at AI. But let's talk little bit about personal computers, you know Santi and I grew up a long time ago, not gonna say the dates, but what was the first PC that you used?

SANTI:
Ohh wow, OK. Well, first PC I used is different from the first PC I owned.

GEORGE:
And I – my answer is different. So let's go down that path.

SANTI:
OK. So the first PC I ever used was an Apple 2E with the green screen. And I used to play this stupid little rabbit game where you have to pull the carrot out of the ground. That was the first PC I ever used. First PC I ever owned was a Commodore 64C, and shortly later I upgraded to the Commodore AMIGA, which was like a 128. Yes and so that –

GEORGE:
Right.

SANTI:
Never on the Tandy. I never owned a Tandy.

GEORGE:
I did not either trust us, right? So I'm very similar, right? We just spoke briefly before this call. We talked a little bit about the Commodore, but my first – the first PC I ever worked on was also an Apple II. It was in high school and we did some basic stuff. We did VisiCalc on it so that was the first business app that I ever used. And then we got a Commodore 64 and I can still remember – at the time, there was – nobody had computers, right?

SANTI:
Command lines, right?

GEORGE:
So I lived in South Jersey. We had to drive up to Philadelphia to go to a place called Bundy Computer, which was in the basement of some little building on a corner. Right? They had a couple of them left. They only got so many a week got shipped in, so we got it brought home, right then went from green screen then to you know full color screen with the ability to do sprites and all kinds of stuff right from the graphic standpoint, but impact, right? I started immediately in high school using it in college and although we advanced technology a little bit and doing papers using – Word Perfect actually. At the time since –

SANTI:
Killed the typewriter.

GEORGE:
So the level of ability to do things in a in a better way and produce better outcome, right? So proximity, productivity hack with that, the advent of the internet. –

SANTI:
So you’ll getta kick out of this, George. Watch this – see if you can remember this? Load, open quotes, asterisks, close quote, comma 8, comma 1 – that loaded the first executable right, on the disk, and then if you said list open quotes, dollar sign, close quotes, comma 8, it gave you the directory.

GEORGE:
On the disc.

SANTI:
Yep, I still remember that.

GEORGE:
And for anybody, anybody younger in the call, they don't know what that means and they're not going to know the next comment which is we had a hole punch and you'd punch the other corner out on the. So then you could flip it over and use. Both sides so.

SANTI:
Use both sides. That's correct, yeah.

GEORGE:
Alright, the next thing was the Internet, right? And I my first real experience was in grad school looking at using mainframes initially using a PS2 and IBM PS2 to do some of my interactions with stuff with school, didn't even have a browser at that time. Towards the tail end of grad school, worked at a financial firm, and that was the first real access to be able to go on the internet. At that time you did things like type in ford.com because it was a name you'd be like, oh, let's say if they have a website. But pretty quickly after that, I was in another grad school program and got a job at the time at KPMG, really at the advent of businesses starting to use the internet. So we built one of the first online gambling sites, which was a live – wheel start spinning, let it start spinning, right, and people were able to wager on it.

SANTI:
Yeah, yeah, yeah.

GEORGE:
We built one of the first ever interactive real estate sites. It was a tremendous game from a productivity standpoint, right, and I expect you experienced some of the same kind stuff –

SANTI:
So before the internet, back to the Commodore, I had a BBS, a bulletin board system and I used to trade games. It was basically over a modem that was this big and it took the entire night to get something sent over and vice versa. And so my first experience with the internet, was just a standard AOL. I basically just signed up for AOL early on a Compaq computer and I started learning HTTP and HTML and the whole 9 yards. But it was a revolutionary time, that's for sure.

GEORGE:
And there was, yeah, it was – I forgot about that right, right, there was AOL and there was Prodigy at the same time that provided a lot of functionality that you see today, right, information repositories, the ability to do Chats, group chats, right, forms, those things.

SANTI:
That's right. That's right.

GEORGE:
What I saw that evolved from an internet standpoint, though, was a number of things. There was a point in time that I worked for a large company that did computing equipment and also did some kind of operational stuff for online wagering. So at the time, the quickest way to do interactions between different systems and this was horse racing. So it was simulcast racing and wagering between racetracks. A lot of the stuff we initially started to do was across the early days of the internet backbone because at the time it was one of the quicker ways to do it and one of the cheaper ways to do it right. We had satellite communication for some things, but mostly video and the wagering systems worked off that background. We know where that led, right? But there were a number of other steps. Smartphones came out, right. They changed the world. You had Hunters initially, you had BlackBerrys, we should do a review at some point – there's a new movie coming out about the BlackBerry. I still have mine somewhere. I've got probably three or four of the old ones. I've got one around.

SANTI:
I had the blue one. I had the blue BlackBerry.

GEORGE:
But that changed the way you interact. You had real time access to people like you have texts today, right, that was the early days of the functionality of the smartphones. And you think about the computing power. You know, I don't even know the actual capability, but my guess is probably the first BlackBerry was more powerful than the Commodore 64. I'd have to think.

SANTI:
For sure, oh of course. Just the processing power alone was.

GEORGE:
Yeah, and say, you know, there's some that are advanced as, as you know, traditional lower end PC's that are in the market. Then we saw cloud, right we saw shared, sort of SaaS, sort of these other kind of functionalities going on from a software delivery model standpoint, but then what really opened up with cloud was the ability to pretty much put any service in the cloud, right, on a shared computing platform, you generally got an increase in availability. You could put other kinds of redundancies in to create a better disaster recovery scenario, right, and that really, really changed things, right. But back to the speed of change, right. PC’s changed – at the time, we thought really quickly, but that was not quick, right? Commodore 64 came out. I don't really know the number of years that it was viable in the market, but it was a couple of years, right? It's not like it's in market for three months and then was outdated. Yeah, the internet comes out, there were constant updates and protocols move fairly quick. Smartphones wicked fast, right, Apple’s on a, you know, yearly refresh now and you see that with other platforms as well. Cloud continuous upgrade. Yeah, right. Uh, but still incremental change over time, but man, AI – the speed – it's just where we were six months ago to where we were three months ago to where we are today, right. And looking, yeah, where are we going to be six months from now?

SANTI:
I love it. Not even close. Not even close.

GEORGE:
Right. Or three months? And that's what's really –

SANTI:
Existing…. That's really it is, this is – and mind blowing stuff.

GEORGE:
But we use it in our everyday work. What we did a couple of months ago has now been transformed already, right, and it's more productive. Now the thing that stood out to me as I sat through some of these sessions today, and we're going to hit on the next slide – on what Microsoft had listed as their top five announcements – and we're not going to go through all five. A couple of these tech. But where are the competitors, right? Where is Google? And if you look at G suite versus Microsoft 365 as a platform, right, a broader enablement platform. Where are they? Right. Yeah, we've heard about Bard. We've heard about some other stuff that they're launching, but wow, the speed at the pace and the speed that things were moving that we saw today on the Microsoft conference. It's just so different than the past, right? Where’s Apple? And where are they?

SANTI:
You know, they'll catch up, they'll catch up.

SANTI:
But they're not leading right now. That's for sure.

GEORGE:
Yeah, for sure. So I'm going to hit the next one and we'll hit on the top five and I'm going to going to play the first couple of these. We're only going to hit on the ones in yellow. I'll start with the note at the bottom and then I'll come back to this, and I'll sort of push some of these to Santi to get a reaction on. But, “the model or the platform right, is not the product.” And this was Kevin Scott, the CTO and EVP of AI at Microsoft, made that statement during the kickoff session today. And we'll come back to that on the last slide that we're going to share, but it's – you got to think through it. It's fairly insightful, right, just because you have a new delivery model or a new tool doesn't mean that it your business is different, if the core of your business can't be effectively enabled by that. And I literally on the fly with Santi, threw together a list of five things will hit on the next slide. But let me go back to the top here we'll hit on the first three on this. So one of the announcements was search grounding in Bing for ChatGPT, you know if you don't know what grounding means, you're sitting there around. Oh, I don't know what that means. So Santi you want to hit on that – What is that? Why does it matter?

SANTI:
Yeah. So grounding is a term that Microsoft has been using a lot lately, especially when we're speaking to AI, but basically it means limiting or restricting your data or the AI to your data. So it's grounding the AI to your specific data. OK, there's different things that Microsoft is rolling out. There's Microsoft Graph, and now they're going to also mention Microsoft Fabric which is taking all the different Microsoft graphs in the organization basically, but grounding is when you take something like a ChatGPT and you narrow the search down to this amount of data. In this case, what they're doing is they're grounding the ChatGPT search results in Bing. And so they're –

GEORGE:
So you got real time, right? So we're no longer lagged by year, year and a half, some unknown sort of lag in the data, right? We've got live real time data. Yeah, I wonder, we, you know, I talked today to Sam Husbands who's on our team has been on the podcast before. He runs our digital for us from an external standpoint and we talked a little bit today about what does this change in search alone, right? How is search different or how do websites operate differently as a use case? Will I a year from now just go to a chat box on a website or have some interactive avatar who I talked to and say well, tell me about your services, right?

SANTI:
Sure, I can see that.

GEORGE:
Is that the interaction? Does the website in its form today not really make any sense?

SANTI:
Yeah, I can see that.

GEORGE:
Uh, so the next two get really interesting, right? So they spoke about Windows Copilot, which is the addition of the ChatGPT enabled tool that effectively an assistant that sits on the side of windows that they gave some basic examples like it helps you – it helps you with the settings on your PC, right? How can I be more productive, and it will give some recommendations around those settings. There's deeper use cases there, but the broader implications of this are the Copilot stack right, so this is going to enable third party use of this, and they give many examples on the – in the conference day and this kick off around how there's some tools already out there are emerging right, that are in beta, but if you want to talk a little bit about those.

SANTI:
Yeah, yeah, yeah. So Copilot in general is going to be – think of it as the reasoning engine that sits between the Microsoft applications, your data and ChatGPT. That's what that is, right. And so, and it does this reasoning, right, using what is known as a Responsible AI framework that Microsoft has, ok. And so this is the reasoning engine. The fact that they're including it in Windows - it's kind of like when they try to roll out Cortana and it didn't really take off. The idea is to help you with your operating system, right, as far as your settings or changing things around or going to a dark theme versus a light theme. Where do I find this feature? How do I turn it on? And so Copilot, keyword Copilot, so you're still driving, leads you through that process. But the Copilot stack is taking that same reasoning engine and saying “hey developers, why don't you go out and develop a copilot for something else? Why don't you create a copilot for this business problem that you're currently having?” And so that is huge, because once – whenever you open up something to the technology community the possibilities are literally endless and that is very smart on their part to say look, here's the stack. Have at it. Go to work and let's see what you come up with.

GEORGE:
And the thing you know, the thing we saw today – we had a conversation earlier today with a CIO. You and I did right, so we talked where you went through some of the basic AI builder stuff, right?

SANTI:
Yeah, that was great. Sure you have.

GEORGE:
You have a technical capability, right, but you're not necessarily an Azure programmer, right?

SANTI:
I'm not a programmer, I'm not a developer. I've been in the IT industry for over 30 years. I know IT. I understand how it works, I know networking. I can put together a computer network. I can put together – I can build computers. I can put together – that kind of thing, but I've never had, I never took a developers course, right. Never did. The closest I got to was some Basic when we had the Commodore 64, especially cuz he had no choice.

GEORGE:
Yeah, you didn't have a choice.

SANTI:
You had no choice. But that was about it. But then here comes Power Platform, right, from Microsoft. They, now granted, they like to say no code or low code. It’s low code. There's no such thing as no code it’s low code, but they don't want to hear that. But that's the fact and it's – but it is low code and that means that the information worker now can go ahead and do development work without knowing code. This is the shift, however here it comes – if you have the Power Platform that can do programming with low code for everybody, what are developers supposed to do? Develop the AI, that's the whole point. That's where the shift is coming, now you're a developer, you're not focused on trying to just develop applications at the code level. No, you're looking at the AI modules you're going to build that's going to make that code even better. So everything is shifting, the information worker is now looking at Power Platform to build AI modules. The developer is now looking at AI to make their code even better and to build better things using the Copilot stack and whatnot. So yeah, this is – well listen, I think you can agree out of all the big revolutionary changes in technology that we just listed, right. This is the biggest in our lifetime. 100 percent.

GEORGE:
Oh absolutely, if it keeps moving right, and it enables everybody effectively to be a programmer. It goes back to the first episode we've recorded. You got to make your own choice or you're going to be a pioneer on this, are you gonna be driving it forward or are you going to be a follower, right?

SANTI:
Or a luddite.

GEORGE:
Right. Or are you going to be a luddite. Who are the bottom? Who are going to fight against it?

SANTI:
Rage against the machine.

GEORGE:
It feels like we're getting past some of that, at least where there was everybody was trying to, you know, identify the bad outcome. So let’s go to the last slide here. This was just a really quick attempt at, you know, what are the new business models? What are some of the things you need to think about? We’ll maybe evolve this over time a little bit more, but it goes back to the point of this is a platform that's there to enable it right or enable AI. It's not necessarily a business model unto itself. Your business needs to be doing something else, right? And this is the secret sauce that's going to enhance it, that if you don't take advantage of it, your competitor is going to or maybe your competitor already has, right. I go back to the – again, we had a conversation with a CIO earlier today and they are at a health services company. They deal with a lot of inbound phone calls that they interact with people, oftentimes repetitive calls about coverage or other service availability or something else related to their health insurance or some other health services. But what he asked us was, could we take a CCaaS platform (so contact center platform) apply something like Copilot. This was before we saw the expansion of this, right, this was earlier this morning, and have that AI look at – if it was Santi calling me on the phone and Santi’s called three times already asking about dental insurance and getting a dental implant. Could the AI see that it was Santi calling, prompt the agent after it goes out and sees all of the other interactions with Santi, all of the related coverage details about that previous service he asked about, and be able to immediately prompt Santi and say “I know you've called about dental implants before is that what you're calling about today?” Yes. “Oh, is it something to do with this right?” And this could be a live interaction with an agent, or it could be a live interaction, effectively with an AI bot, right, either through text or through voice. But what this CIO told us was if I clipped 40 seconds off every call, it would be transformational to his business, right?

SANTI:
Oh, sure, for sure yeah. Listen, and that's the key, right, so what are the things that call centers look for is call resolution. How quickly can you solve that problem? 40 seconds doesn’t sound like much, but in that environment, it adds up. It's a big deal. You know, and think about, you know, to your point what is possible right, in that scenario, the agent would have to go search a database on knowledge base to try and find the answer –

GEORGE:
Right, it could be third party database, it could be their own database.

SANTI:
Absolutely. To have the AI – and this is what I think we're trying to say here – to have the AI be smart enough to know who is calling and what are the possible outcomes or why you do you think this person's calling – this is the third time and this was their interaction before. Why do you think it's called? AI goes ahead proactively researches this, has an output ready for the agent, in a second, right. It's almost like a predictive module where it's actually analyzed and saying I can tell you more than likely they're going to want either this or this so here's the information you need. Yeah, that is huge and I do know that stuff like that is – there are some CCaaS modules out there that kind of do that, but to take it to that level, to where you can, you know, analyze not only – it's a combination of things, it's analyzing multiple knowledge bases and previous call behavior and engagement. By the way, add to that the sentiment, you might as well. AI says, hey, last time he called, he was upset. So let's give him, like, the best treatment ever. And let's get him the answers now. Oh, by the way, and we shaved off 40 seconds off that call. Yes, that's like a win. Then, like, who wouldn't sign up for something like that? Right. And I think all of that is possible to be honest with you.

GEORGE:
And we do – and we see some elements of that today in some CCaaS platforms, right. But we don't see it as deep of a knowledge or a predetermination, right, of this is what Santi may be calling about. It could even pull in other interactions you've had on their website and see that you were looking at an FAQ for a certain topic and have the agent prompt that, right. Again, 10 seconds here, 10 seconds there, all of a sudden you're at 40 seconds across 20-30 agents and you've made a dramatic impact. This thing goes to the – you know we – and again we talked about this in a couple of podcasts, is AI going to take my job. If you are the last person on the fringe of that call center, yeah, your job may not exist, right. There may be some other job for you somewhere else, but yeah, there's some incremental piece when productivity improves that you need fewer workers. It's been like that through that previous curve we looked at for the history of time right, but you need to –

SANTI:
Sure, sure. Yeah.

GEORGE:
You need to make a decision now – are you in or you out right? It is going to be thrust on you. You're going to see Copilot show up in PowerPoint and Word and everything else that you're using every day. But why not proactively start to do it? So let's – I'm going to hit really quickly on this last couple of points and then we'll close this out. But and these are just more thought provoking, right? Santi I went back and forth and looked at this list today. So what is possible, which we talked a lot about and there's a lot of possibilities, right? What's relevant? Do I care? Does it matter? Right, do my customers care, right? What is safe? And there was a lot of comments about safety on the kickoff segment today at Microsoft, not a lot in that on depth and we'll explore some of this a little bit more in some other episodes, but that is that is going to be a key thing that's going to end up driving back to whether businesses put the frameworks in the guiderails in place for this, or does the government do it? What is reliable? So can I rely on this set of services on the back end to be there and be available and be accurate and all those kind of things. And then the final thing from a business standpoint, it sort of goes back to what is possible, is it valuable? Does even matter? Should I invest the money to do this? Is it worth testing it out? Santi and I clearly think it's worth testing. Now this stuff is really inexpensive. Some of the base level licenses you're looking at are in the hundreds of dollars, not the thousands or 10s of thousands to get access to these, right? You need to be testing now. Everything will be completely different six months from now, but being on that journey from today to the future, will make you – will put you in a better position competitively and allow you to make smarter decisions along the way rather than six months from now deciding to jump in and take advantage.

SANTI:
Yeah, of course. And even ChatGPT’s paid version. It's not that expensive. And the whole Midjourney paid version is not that expensive.

GEORGE:
It's not that expensive, yep.

SANTI:
But yeah, we're not talking about thousands of dollars here to use probably what is the most, the most advanced, most mind-blowing technology available to you. I mean, people who’ve been in technology for as long as we have, we would have seen this. I mean, we always knew it was gonna come, but not like this. Right. Like we would have had – and so, and we're talking about peanuts when you really look at the different models and the different charges, which I think is smart. Because if you start to make this stuff too expensive, then you don't make it accessible.

GEORGE:
Totally agree, yeah.

SANTI:
So I think that's a –

GEORGE:
That's part of the, you know, do you regulate or not regulate? People start to close the system off then you're going to end up with regulation, right? But Microsoft clearly is not doing that. Clearly, they are making this – effectively this is not open source, right, but it is an open platform that anybody can access for a really, really minimal cost. So with that, I'm going to turn it over to Santi to do his traditional closeout on the podcast –

SANTI:
So listen folks, we have reached the end of this podcast. One thing that we would really appreciate it –George and I both would appreciate that – is if you go ahead and subscribe on your favorite podcast platform – doesn't matter if it's Spotify or Apple music or maybe it's YouTube. Whatever you prefer, go ahead. We are available in all those platforms, find us subscribe, hit like and share. Let's share the good news. But, you know I could talk about this for another half an hour. You know that, right? I really can, but we can't. We have to stop. So till next time folks, remember to stay connected.

CLOSING VOICEOVER:
Visit www.fusionconnect.com/techunmuted for show notes and more episodes. Thanks for listening.

Episode Credits:

Produced by: Fusion Connect