The Catalyst by Softchoice

The Multi-Cloud Mandate: How Agentic AI Became the Unexpected Answer

• Softchoice • Season 7 • Episode 8

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 26:49

Multi-cloud used to be a dirty word — something that happened to you through mergers, shadow IT, or teams gone rogue with corporate cards. But the walls came down, the standards converged, and best-of-breed finally seemed within reach. Then AI arrived with a whole new layer of complexity.

Or did it?

In this episode, we explore how agentic AI might actually solve the thing that made multi-cloud hard in the first place. Three cloud experts—Jack French from World Wide Technology, Alex Kozaris from Softchoice's AWS practice, and Ron Espinosa from Softchoice's Google Cloud team—break down what's changed, what matters for mid-market teams, and why the "gold record" might finally be possible.

 

Key Takeaways:

• Why 90% of organizations are already multi-cloud (whether they planned to be or not)

• How abstraction layers and platform engineering help smaller teams manage complexity

• What each major cloud does best: AWS for builders, Microsoft for productivity, Google for data/AI

• The compliance curve ball forcing some organizations into multi-cloud for AI governance

• How agentic AI creates "connective tissue" that makes integration problems irrelevant

 

Featuring:

• Jack French, Senior Director of Cloud, World Wide Technology

• Alex Kozaris, Public Cloud Leader for AWS, Softchoice

• Ron Espinosa, Google Cloud Category Director, Softchoice

The Catalyst by Softchoice is the podcast dedicated to exploring the intersection of humans and technology.

Jack:

There used to be that the cloud service providers were very walled off and it was their cloud or nothing. And especially for AWS and AWS fans. So I don't mean this to be, you know, negative, but talking about hybrid or multi used to be bad words, uh, you know, to AWS years ago,

Heather:

bad words, that's what multi-cloud used to be in certain circles, not a strategy. A confession, something that happened to you through mergers and acquisitions, through Shadow it, through that one team that went rogue with a corporate credit card. You didn't choose multi-cloud. You ended up there and then you spent years trying to untangle it, but something changed. The walls came down. Kubernetes became the standard. The big three started supporting each other's protocols. Suddenly multi-cloud wasn't a mess to manage. It was a strategy to pursue.

Alex:

Businesses who wanted to be in the cloud have generally already migrated or decided not to migrate. Uh, now the conversation has changed from to cloud or not to cloud into what's the best solution for this particular business goal.

Heather:

Best solution, best tool for the job. That was the promise, and for about five minutes it felt like we'd finally arrived. And then AI happened. Suddenly there was a whole new layer of complexity. On top of the complexity we just figured out,

Ron:

bar none, everyone's talking about AI and what that means to their business. Now you tie multi-cloud into that and you say, well, now I've gotta do ai and here's an extra layer of complexity. I have multiple clouds running in my business.

Heather:

But here's where the story takes a turn. Because what if the thing that seems to complicate multi-cloud is actually the thing that finally makes it work?

Ron:

Take this command or this query, interpret it and run off and go find this. And I don't care where it lives.

Heather:

I don't care where it lives. That's the promise. That's what's different. Now I. From Softchoice, a worldwide technology company, this is the catalyst. I'm Heather Haskin. This season, we're doing things a bit differently. We're making audio documentaries, real stories from the front lines of it, exploring the challenges of small teams chasing big dreams. Today's episode, what happens when the infrastructure dream you've been chasing for a decade finally arrives, right as everything changes. Again, we're calling it the Multi-Cloud Mandate Act one already here. Let's start with where we actually are, not where we're going, where we are. Jack French is a senior director of cloud at our parent company, worldwide Technology. In his day-to-day, Jack lives and breathes cloud strategies, and what he's seeing isn't a trend toward multi-cloud. It's a recognition that multi-cloud is already the reality.

Jack:

Almost all organizations, at least that we're working with in the enterprise space at least, are deployed and have a multi-cloud strategy. So it's no longer the days of, you know, cloud first and trying to operationalize inside of one public cloud, really, whether it's intentional or through mergers and acquisitions or other means. Almost all large enterprises have found themselves in a scenario where they're adopting a multi-cloud strategy,

Heather:

whether it's intentional or through mergers and acquisitions or other means. That's the key phrase. Multi-cloud isn't something most organizations chose. It's something that happened to them and now they're trying to figure out what to do about it.

Jack:

89% or 90% of organizations are leveraging multi-cloud for one reason or another,

Heather:

90%. That's not early adoption. That's not a trend. That's the baseline. Alex Ceros leads the AWS practice at Softchoice. He's been a software developer, a systems administrator, a DevOps engineer, and he's watched the conversation around cloud shift dramatically over the past decade.

Alex:

10 years ago, there was much more of a philosophy of lifting and shifting. So just take all of the virtual machines or or bare metal servers in your data center, move them into the cloud and then, you know, great. Things will happen. We'll save money. We'll, you know, get these benefits, better security. But right now, everyone who wants to be in the cloud has moved. Those who are still thinking about it now, want a little bit more of a concrete relationship between what their specific business needs are, both now and into the future. And then how Cloud can help them achieve their goals.

Heather:

The question changed. It used to be, should we move to the cloud? Now it's which cloud is best for this particular thing we're trying to do? And that's where it gets interesting because different clouds have different strengths and different parts of your organization have different priorities.

Alex:

I have actually seen this with businesses who are making the decision of which cloud to move into. And the IT department wants one cloud because it suits their need. An innovation group who's interested in AI wants a different cloud because it solves their use case. And then the CEO and CIO might want a different cloud because it provides, uh, what they see to be the, uh, best value for money.

Heather:

Three departments, three priorities, three different answers to the same question. If you've ever been in that room watching smart people argue past each other about infrastructure. You know how this usually ends, somebody wins, everybody else compromises, and you spend the next few years living with a decision that made half the company unhappy. But what if you didn't have to choose? Here's what actually enabled multi-cloud to become practical. It wasn't a single breakthrough. It was a series of shifts, some technical, some philosophical that added up to a fundamentally different landscape.

Alex:

In the past, um, cloud providers tended to innovate in silos, and the idea was to build the best service they could. But in the end, they found that customers didn't wanna lock themselves into a particular, uh, technology. So they all decided to support Kubernetes with a managed service. Azure. And Google copied the S3 protocol to make sure that any applications built for AWS could also run on Azure and, uh, and Google.

Heather:

Kubernetes, the S3 protocol. These aren't just technical standards, they're the building blocks of portability. When everyone speaks the same language, moving between clouds becomes possible.

Jack:

I think over the last several years you've got, you know, platforms like. Snowflake or Mongo. And then you've got even the openness of the Microsoft, Amazon, and Googles of the world of creating these interconnected networks that now are aware of each other and allow, you know, communication much smoother. You've got the advancements of containers and the management of Kubernetes and these different container services that allow you to create portability and applications now that make it much easier to move between, you know, either public clouds or even your own hybrid private environment.

Heather:

But why did the vendors open up? These are competitors. They want your business. Why would AWS make it easier for you to also use Azure?

Jack:

The more boxed off you are, or closed you are, the higher likelihood you'll be as the one cloud that's left out.

Heather:

Fear of being left out. Fomo. When 90% of organizations are already multi-cloud. The vendor that refuses to play nice isn't protecting their market share. They're shrinking it.

Alex:

I don't necessarily think they're playing nice with each other as much as they are meeting the customer where they are. So where there is demand for customers to be able to run in multi-cloud, they will support that and provide that service.

Heather:

Meeting the customer where they are. That's the shift. Not all truism pragmatism, so that's the enterprise reality. Multi-cloud isn't coming. It's here. The vendors have opened up. The standards have converged. The dream of best of breed using the right tool for the right job is finally within reach. But what about smaller teams? What about the mid-market IT leader who doesn't have a dedicated cloud center of excellence or a seven figure consulting budget? That's where things get more complicated. Act two, the mid-market reality. Here's the thing about multi-cloud, the benefits scale. The complexity doesn't. If you're a Fortune 100 company with a massive IT organization and dedicated teams for each cloud platform, multi-cloud is manageable. You've got the talent, you've got the budget, you've got the headcount to absorb the complexity. But if you're a mid-market company with a Lean IT team, the math is different.

Jack:

When you get into small SMB or mid-market teams, they're going to lack those larger teams that allow them to have that diversity of talent across multiple cloud providers. And so in those scenarios, it's even more important to have well-defined policies and practices. A really strong operating model around your cloud strategy so that you are addressing those risks and complexities that multi-cloud introduces with automation and with almost like an abstraction layer.

Heather:

An abstraction layer. That's the key concept for smaller teams. Instead of managing three different clouds. With three different interfaces and three different sets of expertise, you create a single control plane that sits above all of them.

Jack:

By leveraging platform engineering and creating this abstraction layer or a control plane type of environment that allows you to deploy resources from a single location and ensure that those resources. Have consistent tagging strategies and security policies and all the governance that's required around that. Creating that, you know, abstraction layer allows you to move faster and make it so much easier to move resources into the cloud too, and building that. Properly from the beginning and defining that operating model from the start allows you to move much faster later and keep you from, again, ending up on the news and potentially damaging your brand by some security risk or, or breach.

Heather:

Deploy, once, run anywhere consistent security, consistent governance. That's the promise. Of the abstraction layer. But Jack is careful to note this isn't simple. It requires upfront investment in platform engineering. It requires defining your operating model before you start building for mid-market teams, though, there's another angle worth considering. It's not just about managing complexity, it's about leverage.

Jack:

In enterprise, you could have billions of dollars of committed cloud spend each year, and you know that volume gives you some pretty significant. Discounting leverage. And so for the mid-market, or maybe you're talking about tens to hundreds of millions of dollars versus billions, how do you still create leverage between the cloud service providers to make sure that you're getting the best discount? It could be that you've got the ability to leave one of the cloud providers at any point. So creating that portability to creating. The multi-cloud environment in your mid-market or smaller enterprise environment could tell in Azure, Hey, I, I've got environments spun up with production workloads in AWS. At any point I could shift this over, or vice versa.

Heather:

Negotiating leverage. If you're a mid-market company without billions in cloud spend, your discount conversations are different. But if you can credibly threaten to move workloads between providers, that changes the dynamic. Now if you are going to pursue a multi-cloud strategy, it helps to understand what each platform is actually good at because they're not interchangeable. Each one has a lane, Alex frames AWS as the builder's platform.

Alex:

So I've always. Seeing AWS is more of a builders platform. They're just constantly adding. I think they've got over a hundred now, uh, different services. And the goal is to give developers, engineers, uh, business leaders as many ways as possible to build efficiently and address. Their use cases. Uh, Microsoft on the other hand, they have a very strong market with Windows and SQL Server and in the workplace productivity area,

Heather:

AWS for building Microsoft for productivity and the existing Windows ecosystem and Google. Ron Espinoza leads Google Cloud at Softchoice. He's clear about where Google has planted its flag.

Ron:

Google differentiated around data and ai. And I don't think anybody would argue that Google, when you, if you're thinking about doing data and AI in the cloud, you should be thinking about Google. It's just where they've made their mark

Heather:

data and ai. That's Google's lane. And here's what's interesting. The vendors have become comfortable enough with multi-cloud that they're actually making their specialized services available across platforms.

Ron:

If I'm an Azure customer and I've got an Azure stack and I've got an Azure shop that I'm running, all my people are trained on Azure, but I want BigQuery to run inside of Azure. Wouldn't it make sense if I could. Well, guess what? Google became amenable to that and made that a reality.

Heather:

BigQuery inside Azure. That's the kind of interoperability that was unthinkable five years ago. The walls have really come down, so you've got a clearer picture now. Multi-cloud is already the reality for most organizations. The vendors have opened up, the standards have converged. Mid-market teams can pursue best of breed strategies if they invest in the right operating model and abstraction layer, and then AI arrived and everything got more complicated. Or did it act three? The connective tissue. Here's where things stood at the beginning of 2024. Multi-cloud was finally practical. The technology had caught up with the vision. IT leaders could realistically pursue best of breed strategies, and then AI exploded and suddenly there was a whole new set of questions on top of all the questions they just figured out.

Ron:

Bar none. Everyone's talking about ai. And what that means to their business. And as we talked about in our last segment, some folks know what that means. Most don't. Some folks have a plan, most don't. And some folks are executing, most are not. And so now you tie multi-cloud into that and you say, well, now I've gotta do ai and here's an extra layer of complexity. I have multiple clouds running in my business, which most companies do. And they're not asking, how do they tie them together? They're saying, how do I make AI work across them?

Heather:

How do I make AI work across multiple clouds? That's the question. Keeping it leaders up at night and it's getting more complicated. HIT compliance. Jack sees organizations being forced into multi-cloud, specifically because of AI governance requirements.

Jack:

When you look at organizations that have data sovereignty requirements, or compliance and governance, think of even like our federal government or you know, the public sector space. You may find that, you know, some services. Have like high side accreditation or FedRAMP certification while others don't. And so in a scenario like that, you may find yourself using some resources that are certified and have all the governance and sovereignty that you need, but you've gotta pull in another service from another public cloud provider because it has a clearance or a capability. That the other cloud does not.

Heather:

And here's a specific example of that playing out right now

Jack:

in a scenario where an organization may be highly committed to like M 365 for their productivity and office stack, but they need, you know, Gemini Enterprise because it's got a. Accreditations or certification that Microsoft doesn't in that scenario, uh, that's a situation where, yeah, you, you may be forced to use a multi-cloud environment.

Heather:

Gemini has accreditations that co-pilot doesn't. So even if you're all in on Microsoft for everything else, you might need Google for ai. That's multi-cloud driven by compliance, not choice. Meanwhile, the cloud providers are making big bets on where AI is heading. Alex was at AWS Reinvent 2025, and he just noticed a clear strategic shift.

Alex:

Werner Vogels, the CTO of Amazon. He's a brilliant leader and engineer, and he described what he called the Renaissance developer. So developers of the future, according to him, will need multi-domain expertise to think strategically about solving business problems and then leverage AI to execute on those. And so that's why AWS has gone all in on agentic ai.

Heather:

A agentic ai, that's the hot phrase. AI systems that don't just answer questions, they take action, they execute tasks, they work autonomously. And Alex has a theory about why AWS is betting so heavily on this direction.

Alex:

AWS kind of, in my opinion, fell behind during the generative AI wave and bedrock, which is their tool, is excellent in terms of capabilities. But Microsoft has that integration with office to build off of, and Google's service portfolio includes Gmail, G Suite, Android devices. Um, AWS doesn't have that same connection directly to consumers. It makes sense that they would, uh, pick a agentic AI to push on.

Heather:

So here's the landscape. Microsoft has the productivity suite. Google has the data and AI heritage. AWS is betting on ag agentic ai, and the builder community. Three different strategies, three different strengths. Which brings us to the question at the heart of this episode. Does AG agentic AI actually require multi-cloud? Alex isn't so sure.

Alex:

I don't necessarily think that going multi-cloud is a requirement in order to be successful at leveraging ag agentic ai. However, what multi-cloud strategies make possible is the ability to use the absolute best tool for the job, for the business case that you're trying to solve.

Heather:

Not required, but enabling multi-cloud lets you use the best tool for each job. That's been the promise all along, but Ron sees something different happening, something that might flip the whole conversation on its head.

Ron:

Now we're talking about with the advent of Ag agentic ai and generative ai and and AI in general, and the widespread use of it, being able to point an assistant or an agent at a problem using natural language and letting it learn. Through programming and, and through natural language and through lots of other things that we won't cover in this call that, hey, take this command or this query, interpret it and run off and go find this. And I don't care where it lives.

Heather:

I don't care where it lives. Sit with that for a second, because that's not just a feature description. That's a fundamental shift in how we think about multi-cloud for years. The challenge of multi-cloud was integration. How do you get data from Azure to talk to applications in AWS? How do you maintain consistency across platforms? How do you avoid the egress fees and the API translations and the endless complexity? What if you didn't have to solve that problem at all? What if you just pointed an AI agent at it?

Ron:

I've got Application A lives in Azure, application B lives in Google application. C lives in AWS. Together. They give me the gold record, if you will. The agent goes and finds that, and now I don't have to merge all that data together. I don't have to integrate clouds, I don't have to pay egress fees and and worry about ingress and how data travels from one cloud to another, I just deploy an agent and it goes and ferrets all that information out and pulls it back.

Heather:

The gold record. Your data scattered across three different clouds stitched together, not by painful integrations, but by an AI agent that just goes and gets it.

Ron:

Now with agent to agent protocols where you can build agents that talk to agents and agents of commerce or agent to payment type of protocols, you're able to have a real commercial use for this, whereby we're eliminating the need for people to be involved. And think about all the levers that have to be pulled for somebody to go query the Azure database, pull all the information out, and then somebody else goes and does that in AWS. Somebody else does that in Google. And what if, because this agent could talk to that agent, any cloud and any microservice can talk to any other one immediately with very little development and very little development cost, it changes the game entirely.

Heather:

Any cloud and any microservice can talk to any other one immediately with very little development cost. That's not incremental improvement. That's a different paradigm. Instead of building integrations, you deploy agents instead of managing complexity, you abstract it away.

Ron:

There's a connective tissue that, that an agentic layer can add that just hasn't been present to date. Certainly not at this speed and and efficacy.

Heather:

Connective tissue, that's Ron's phrase, the agentic layer as the thing that finally makes multi-cloud work, not by solving the integration problem. By making it irrelevant, and here's what makes this different from all the other next big things in it. It's not theoretical, it's not five years away. Ron is seeing it in practice right now.

Ron:

This is a real use case we had with a customer. They wanted to figure out how they could use their existing ticketing system, which was meant for customer trouble tickets. They had used it to onboard. Employees. It's not really the right use of it, but they figured out how to do it. In order to shift off of that, they were looking at going to a major application like a ServiceNow, for instance. That's a heavy investment. What if you didn't have to do that anymore? What if you said to the agent, you created an agent that went and found that information and created the onboarding program for you?

Heather:

Instead of a major platform investment, deploy an agent instead of ripping and replacing, build a layer on top that goes and finds what you need. So where does this leave us? Multi-cloud went from dirty secret to industry standard. The vendors opened up the standards converged, and just when we figured out how to manage the complexity, AI arrived and seemed to add another layer on top. But here's the twist. Agentic AI might not just benefit from multi-cloud. It might actually solve the thing that made multi-cloud hard in the first place. The agents don't care where your data lives, they just go get it.

Jack:

I don't

Ron:

know that there's an impetus to do it. I think that the key message that I would deliver right now is that it's not pie in the sky. It's not 10 years from now. It's not five years from now. It's not three years from now. It's now, it's available now. I would encourage a strong investigation, a serious dedicated investigation if it's right for them. Uh, and I would look at how easy it is. Don't think that it's super expensive to do. It doesn't have to be.

Heather:

It's now, that's the message. This isn't a future state conversation. The technology exists. The question is whether you're ready to use it, but, and this is important, that doesn't mean you should rush in without a plan. The fundamentals still matter.

Alex:

One extremely important, yet often overlooked aspects of a cloud migration initiative is what are our business goals, right? Where do we want to be in one year, two years, three years, and so on with our technology? While multi-cloud has gotten a lot easier, it's not something that we should attack by saying, let's just build all clouds. Let's just spread our workloads on all of them. That's not the right solution either, because it's gonna create a lot of work that may not lead to a return on investment. By understanding your goals, you can work backwards from that and make the right choice, uh, for you.

Heather:

Start with business goals, work backwards. That's the consistent advice from everyone we talk to, and the operating model, the people, the processes, the governance. That still matters too.

Jack:

If you haven't defined how your organization is gonna adopt it, why they're adopting it, and really have aligned the people and process, then the technology will fail every time. So I think it's critical to. You know, answer those questions. Why am I doing this? Does it actually make sense for me to go multi-cloud? And if so, okay, how am I gonna do this? And how am I gonna align my operations, my people, my process, the policies and governance and everything else that's required to adopt this type of framework?

Heather:

If you are an IT leader listening to this and think. We're already multi-cloud. We're already feeling the AI pressure, and we don't have time to figure all this out ourselves. That's exactly what Softchoice helps with,

Ron:

and where Softchoice can help is to slow you down so you can move faster. Is it slow? Is smooth and smooth is fast. That's where we want to be. Let's take a second to really understand what we're trying to do. What is it that's really driving your business? What's your top three strategic initiatives? Let's look at those and let's back into the technology decisions.

Heather:

Slow is smooth. Smooth is fast. Start with what you're actually trying to accomplish. Then figure out which clouds, which tools, which agents can get you there. If that's a conversation you want to have, visit softchoice.com to get started. The Catalyst was reported and produced by Tobin Dalrimple and the team at Pilgrim Content Editing by Ryan Clark With support from Philippe Dimas, Joseph Byer, and the marketing team at Softchoice. Special thanks to Jack French, Alex Kre and Ron Espinosa for sharing their expertise.