Subscribe
Share
Share
Embed
AI First with Adam and Andy: Inspiring Business Leaders to Make AI First Moves is a dynamic podcast focused on the unprecedented potential of AI and how business leaders can harness it to transform their companies. Each episode dives into real-world examples of AI deployments, the "holy shit" moments where AI changes everything, and the steps leaders need to take to stay ahead. It’s bold, actionable, and emphasizes the exponential acceleration of AI, inspiring CEOs to make AI-first moves before they fall behind.
Andy Sack (00:00)
As a CEO, you cannot wait to start moving and your organization.
towards being AI first. Speed matters and adaptability and portability matter. As a result, I, you know, I'd be stronger. I'd be like, absolutely. You should have a team and you should probably have a team of 50 people. You should pay for extra licenses and you need at least one, two or three, of the LLMs. So you can be learning in, in that way with this group.
what works for your organization from a security standpoint, a governance standpoint, and a tool standpoint.
This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI innovation and business. I'm Andy Sack and alongside my co-host, Adam Brotman, each episode we bring you candid conversation with business leaders, transforming their businesses with AI. No fluff, just real talk, actionable use cases and insights for you.
Yeah, so let me tee up. I mean, broadly, the topic of the day is what is our advice and what is the playbook for executives aspiring to be AI first at their company that are dealing with Microsoft and Microsoft Copilot? And I'll illustrate the problem or the challenge because we had a client conversation
yesterday that brought this to light.
As you know, I spent seven years working directly with Satya Nadelle and Microsoft. So my, understand both the organization and I'm partial to the organization and think Satya's doing a great job. That disclaimer noted, we're having a conversation with eight executives at a hundred year old company that has 600 plus employees. They're a Microsoft shop and they're aspiring to
be AI first and where we demoed for them two tools, demoed Lovable and Claude Cowork. And the executives were blown away by both the tools as well as this concept of meta-prompting. And their IT person was like, where we use Copilot, we can't do all of that and the data is not secure, what should we do?
That's the problem that we see over and over again in the market. And we thought it was worth a mini episode. So I posed the problem, Adam, you want to either add some problem, also just offer your perspective on what your advice is to enterprise.
Adam Brotman (02:46)
I mean, I think you frame the problem. I mean, I think there's like thousands of companies, if not hundreds of thousands of companies that are dealing with exactly what you just said. And they're hearing about all these other tools and all this amazing stuff. And they're Microsoft shops or co-pilot shops. And so what do you do? And I think we've got some good ideas and advice to give to companies in that situation.
Andy Sack (03:11)
Yeah,
so go ahead, dive in. Let's give the playbook.
Adam Brotman (03:14)
Well, so for starters, let's start with the fact that, number one, as an enterprise, you should be careful and responsible with what data and information you are giving to an AI system, any kind of AI system for any reason. Because, I mean, you're careful with it with everything else in technology. Why shouldn't you do that with AI? And that's absolutely true. So let's start there. And if you're a Microsoft shop,
and your IT department's like, the only approved system is Co-Pilot for AI use. Like, okay, at least you know it's safe. I mean, you should, you're probably reasonably secure that that's safe. That's good. But the problem I have with that is why couldn't you add
another couple of systems that have safe versions, either literally called enterprise versions or business versions where the models aren't training on your uploads in your chats, where you know that they're like, I'll call it SOX compliant if that's necessary, et cetera. like Microsoft does not have a monopoly on security and confidentiality and data compliance. so my first reaction would be like, fine, be a Microsoft shop.
You know, Copilot has a lot of redeeming qualities. It's a perfectly good system in a lot of ways, but it is not, in my opinion, as powerful and as good as ChatGPT and Claude for a lot of the things that people want to do with AI right now. And that's putting aside ancillary systems like Perplexity and Lovable and Notebook LM, et cetera. So the short answer, and I'm curious to get your response, in my opinion, would be fine. Have an AI use policy.
Let your IT department be an important part of that task force that designates what's allowed to be used. But don't just limit it to co-pilot in that case. Have a couple of people, maybe it's 10, maybe it's 2, maybe it's 20, that are authorized and know what they're doing and give them access to the best tools. Claude 4.6, ChatGPT 5.2 in a secure way, meaning
in a enterprise grade or business grade way so you know you're being just as responsible as you are if you're using Co-Pilot, in my opinion. And maybe even expand that to some of the Google products and Gemini products like Notebook LM, and Gemini. Even Gamma AI and Lovable and these guys, they all have, they read your terms of services. Don't think, my personal opinion, maybe last one is, don't let your IT department,
Andy Sack (05:26)
.
Adam Brotman (05:47)
Jedi mind trick you into thinking that that's the only path. Like you can be responsible and yet still have a small group of people that are using multiple tools. And you and I always tell people, using multiple tools is the way. Like you want to be there anyways. It may be too... Sorry. Yeah.
Andy Sack (05:47)
Look. Yeah, alright, let me let me let me me jump in here.
So I think look like. In an AI first world. You are going to be like the proliferation of tools is happening.
and it's accelerating. And we also say AI first companies start with AI first leaders. This podcast is, we're really speaking to the CEOs in this instance. And I think this is a moment where leadership is required because most of enterprise, Microsoft owns the enterprise. Like they've had a beachhead, they bundle their products, they've done it incredibly well for 40 plus years at this point.
AI and they're going to do the same with AI. They have a great relationship with Chat GPT and open AI. But getting the LLM into the products of Microsoft has organizational challenges at Microsoft, which I know very well, and it's going to delay the adaptability of the enterprise customer.
As a CEO, you cannot wait to start moving and your organization.
towards being AI first. Speed matters and adaptability and portability matter. As a result, I, you know, I'd be stronger. I'd be like, absolutely. You should have a team and you should probably have a team of 50 people. You should pay for extra licenses and you need at least one, two or three, of the LLMs. So you can be learning in, in that way with this group.
what works for your organization from a security standpoint, a governance standpoint, and a tool standpoint.
Today, I'm toggling between really the AI native, the AI first LLMs, really Claude and OpenAI. Gemini, like Google Gemini, amazing. They have some of the same issues that Microsoft does. Those organizations have huge scale and should, you know, everyone's expecting Google to win the
l l race and yet it's slower than both quad and open ai not by a little by by a lot and so i would be more strong and i would lead and the critical piece is this cannot be a case where it wags that the tail wags the dock
Adam Brotman (08:22)
Yeah, and it's worth saying, I follow you, man. It's worth saying, adding a little color to what we're saying. What we are not saying to enterprises in this podcast and in general is don't worry about security, don't worry about the risks. There are a lot of risks. We tell our clients every day, just because you've got an AI use policy and you've got a task force and you've got the right enterprise grade versions and stuff, your people still need to be educated.
and they still need to understand what's going on. You don't want somebody just turning on Claude Cowork and then giving it access to all your systems using computer use. There are things with agents, for example, you gotta be really careful of at work, even if you're using the right versions of things. So I'm a big fan of being responsible and careful, but there's a middle ground here, like you were just saying, where it's like, don't be
don't just be told, I don't think as a CEO by your IT department, like, no, there's no other safe way to do this and just like staying only with copilot. Like it's like, no, I would, I would ask them to say that's not true. There's gotta, you know, fortune 100 companies that are using Gemini and open AI and anthropic. So like they, there must be a responsible way to do that as well. And the devil's in the details. Like people need to understand that a,
Don't get backed into a corner. But B, you still have to know what you're doing. So when we're saying to people, don't just accept one answer, we're not saying don't be responsible. And I'll add one other last point, which I think is really important, which is that if you're doing this right and you've got, let's say you're a co-pilot shot, you got co-pilot, great. Co-pilots eventually, like you said at the beginning of this podcast, I would never bet against Satya. Like I believe
that over and same with Apple, but like this, we're talking about Microsoft. Like I wouldn't bet against these companies. they are run by really smart people. They might have cruft and bureaucracy and be slow and this might be not happening fast enough. But I believe that Microsoft will figure this out in a way where we're not having conversations about like why, why, what do you do? Yeah.
Andy Sack (10:35)
But can Adam, can I interrupt you for a moment on this? Because
I think the challenge for our clients is that I believe what you just said. I believe Microsoft and Microsoft are going to get there. It might take 18 months to two years for Microsoft to get there. And this next two year period.
is critical if you want to be AI first and start the transformation that the human change management, the transformation that's required, you cannot wait for Microsoft Copilot to actually catch up with the AI native LLMs. That's my point of view because the business, your business is going to be affected or could be affected.
Adam Brotman (11:09)
No.
I'm really glad you said that. You're right. To your point, if it took two years, to your point, the good thing, they're listening to you. And I agree with that. And whatever amount of time it takes, we're telling people, don't wait. Start using Claude and OpenAI or whatever in a responsible way. That's kind of our point here. Actually, a group of people, whatever size, that are proficient in a responsible way using multiple tools. And the key thing is,
they're going to be sharing with other people, maybe hopefully the whole organization, what they're doing, then that's the only way the organization's ever going to know what to use in Co-Pilot someday when it does get to the place it needs to get to. So it's all about getting the organization understanding what's possible now and not waiting to your point, but staying responsible. then at some point, Microsoft will catch up and it'll all
be okay.
Andy Sack (12:09)
Here's the thing I made. As I said, I'm partial to Satya at Microsoft, so huge fan. Something happened in January of 2026 and we can feel the exponential curve ourselves and and it's going to accelerate in the next two years. The to illustrate the point with the client.
their business is already being affected by AI in 25. It's going to be more affected by AI in 26 and 27. What we're talking, their processes, which ultimately involve IT and security and ultimately involve, and tooling, also involve IP, also involves organizational processes and workflow.
as this acceleration happens along that as the alarms accelerate and code the LLM start writing more and more code faster and faster. It's going to outpace the ability for that business to adopt and that the impediments and why you need to start yesterday. You need portability. You need you basically need to ignore ROI and you need multiple tools so you can actually first hand. Work with.
the LLMs to figure out what works for your organization and take advantage of the tools as opposed to being on your heels two years from now.
from my mouth to God's ears.
Adam Brotman (13:30)
That's right. That's right. just to sum it up, we should make sure everyone understands that what Andy said. Don't wait. This stuff, you said things are accelerating. It's not like a year. It's going to be like in the next six months. Yeah.
Andy Sack (13:49)
Week to week, it's week
to week.
Adam Brotman (13:52)
Next six months, things are happening and you at least need to be experimenting in a responsible way and understanding what's happening. And you say, ignore ROI, I what you mean. The truth of matter is, it's not that expensive to implement a few other tools in the hands of some portion of your workforce. so we're not talking about like spending an arm and a leg without knowing you're doing. This is actually about, it's not even about money. It's about like just being smartly
being able to see what's happening with these frontier models, whether or not held back by any sort of defensiveness. And you can be responsible with them. So hopefully everyone that listens to this kind of understands what we're saying.
Andy Sack (14:37)
It's why we're partners.
don't mean to be stupid and say ignore ROI. And yet we encounter with our clients ROI. They seek ROI too early when dealing with AI. And you've got to experiment and move and figure it out what works for your company.
Adam Brotman (14:54)
So can I give a real world example? I know this is a mini episode, but we like real world examples. OK, so I was talking to one of our clients today. They are actually a Microsoft shop, but they also have a ChatGPT enterprise license as well. So they're kind of doing what we're saying. They were actually implementing a change management on one of their enterprise systems that
Andy Sack (14:57)
Yeah, and then let's close it out.
Adam Brotman (15:20)
had nothing to with AI, but was a really important system that faced all their salespeople. And they were having to deal with an influx of sale internal support tickets that were going on. And it was interesting because their systems of record were not able to keep up with what they needed. And yet, in listening to what they were saying, I'm like, ⁓ you could just use Claude, for example, or Chat GPT and create a bot for this. And this is what you could do.
And I talking to the CEO and they're like, yeah, I wonder why we're not doing that. Even though this is like a really AI proficient organization. And I think it's a good example of how six months ago, I would have been much more tentative about saying to them, yeah, why don't you just take all these records and take all these transcripts and put it over here, create a knowledge base, create a bot, solve your problem in the next 48 hours. And by the way, the ROI would have been off the charts on there. It wouldn't have cost them anything to do what I'm describing.
Andy Sack (16:00)
Okay. ⁓
Adam Brotman (16:15)
taking advantage
of the general purpose problem solving nature that we know is now possible. And instead, they were sort of stuck working with legacy systems and manual processes. And I think it's a good example of how far these systems have come just in the last few months. ROI, it's going to go from, I wonder what the ROI is to, my god, I can't believe I wasn't able to solve this problem and that problem because I wasn't able to move fast enough.
Andy Sack (16:31)
Yeah.
Adam Brotman (16:42)
and how these other companies are doing this and what do I do? And the advice we're giving, whether it be co-pilot with augmented with other systems or not is understand that the ROI is here now. Like this stuff is able to do things that you in your mind, if you're reading articles about MIT studies and all this other BS from four or five months ago, like you have to understand these systems have greatly accelerated just in the last few months.
And you got to get a group of people in your organization to understand that because as these problems come up, you got to be able to take advantage of it. And the ROI is going to be off the chart.
Andy Sack (17:18)
It touches on a topic which we could probably cover.
is really the organizational changes that are gonna need to happen now. Like, we talk about one title, which was the AI officer, that's insufficient.
AI is fundamentally going to change departments, job titles, how people work together. And we can cover that in another episode. I think that'd be pretty interesting. With that, let's close out this mini episode. Thanks to you, our audience, for listening to AI First with Adam and Andy. For more resources on how to become AI First, you can go to our website, forum3.com, download case studies, research briefing, executive summaries, and join our email list.
We also invite you to join our AI First community, a curated hub and network for leaders turning AI hype into reality. We truly believe you can't over invest in your AI learning. Onward.