Microsoft is positioning itself as the platform of choice

POV by Steve Rattacasa, Katalyst CTO

As a CTO, my mind sometimes does a lightning fast review of all the latest tech issues I need to consider – like touring a digital fairground before diving in.

And what I see at the moment is a profound change in outlook when approaching AI use.

Where the focus used to be firmly on don’t let hackers steal your data, I’m observing a new issue surfacing much more often this year – the importance of not allowing AI tools to absorb your operational know-how, workflows, and feedback loops.

Put bluntly, the thinking is: Why pay the tool to give away your proprietary information? It doesn’t make business sense. And organizations are beginning to realize this.

It’s a different kind of security problem.

Satya Nadella, Chairman and CEO of Microsoft, is clearly thinking along the same lines.

If you missed his recent piece, the core idea is that AI can create a “reverse” version of the old information paradox. In the new version, the business user gives the AI system valuable proprietary context in order to get good results from the tool – but the platform may learn more from that usage than the user does. And disseminate it to other users who ask the AI for information.

If that’s true, then the discussion changes.

It makes choice of model and platform rather more important than raw AI capabilities. Not AI leaderboard status but how your business info will be used. Is it secure?

Obviously, Microsoft (via Nadella) is likely claiming the high ground here already.

It’s recognized that an organization’s context supercharges their AI results and becomes their own intellectual property/differentiation. And Microsoft is also the platform in which much of that context already lives.

In other words, Microsoft is positioning itself as the platform where you can maintain your data and context privacy and still maintain model choice and flexibility (I’ll get back to that later).

Making things easy to integrate, secure, and govern gives Microsoft a “data gravity” advantage.

But I’ve seen this coming in our work at Katalyst.

To me, Sovereign AI is really about keeping ownership of your organization’s knowledge, not simply choosing the best-performing model.

The Issue of Trust and Data Privacy

Data privacy and trust is increasingly important as organizations look to maintain control over their data, AI models, and long-term learning.

Trust in the platform to be secure. Trust in the AI to be isolated, containing all the business’s learnings within the business environment. Not broadcasting them.

I believe our future customers will be choosing AI vendors based on security audits, not benchmark scores.

After all, every prompt, correction, and workflow used with AI can teach the model how the business works – which means the information provider may be giving away competitive advantage while paying for the tool!

My job is to make sure our AI tools improve our service without learning and leaking the client’s secrets, workflows, or hard-won operational know-how.

For that, we need AI to be attached to an already trusted platform.

That’s how I noticed Microsoft has been shifting towards a more “sovereign advantage” message of late. Cisco and other vendors are also focusing on sovereign AI (meaning you control it, it’s your data, your models, etc.) but more from a hardware perspective.

AI produces business intelligence faster than ever. But the business risk now is not just AI cost or accuracy, but leakage of the very operational intelligence that makes the service valuable.

Microsoft has a habit of using its massive distribution to catch up remarkably quickly when the market shifts.

So, at the risk of sounding bullish, Microsoft doesn’t have to convince customers to adopt a brand-new AI operating model. It can attach AI to the systems many teams already rely on.

They already have customer data, security tooling, and many of the foundational pieces organizations need to successfully adopt AI.

And as I indicated, they’ve been embracing a flexible approach to AI model providers.

It’s a highly flexible, multi-model AI strategy. Rather than relying solely on OpenAI, Microsoft offers an extensive ecosystem of third-party, open-source, and internally developed models across its cloud and enterprise platforms.

Taken together, that’s why I see Microsoft as a likely major winner. Because of its platform.

But is this enough to become the premier data privacy platform businesses now need?

Will Microsoft Win The Data Privacy Battleground?

I can’t forecast how this will play out.

What I can say is that Katalyst is vendor agnostic, and as Katalyst’s CTO, I have to look at infrastructure and risk rather than just the technology trends.

Managed service firms often have access to the most valuable part of a client’s business: incident patterns, ticket resolution habits, infrastructure decisions, and service preferences.

So what I’m saying here is that the winner may be the company that already owns the work surface, the data path, and the trust layer.

And Microsoft is seen as strong at the moment, so it could capture outsized value even if other firms build the most advanced models.

For example, one clear current advantage is they offer centralized governance. Admins and enterprise users can opt to plug in independent AI models based on their distinct needs regarding costs, data residency, and privacy constraints.

I see Microsoft positioning itself as the platform where you can maintain your data and context privacy and still maintain model choice and flexibility.

But harking back to my initial comment about not letting hackers steal your data, taking the option to NOT feed your proprietary business information and learning into a model that AI scrapers can then benefit from – well, it sounds to me like the next version of competitive advantage.

It also changes where I think leadership attention should be focused.

Before choosing a model, be absolutely clear about the business problem you’re trying to solve. Only then can you decide what information the AI genuinely needs access to.

Demos are easy. Real integration is hard.

And forcing everyone to use AI without the right controls can be counterproductive to data privacy and in-house security.

Whether Microsoft emerges as the leader in this new arena is by the way. Giving away your hard-earned competitive advantage seems madness to me.

Why Katalyst Prioritizes Data Privacy

We work with you as your partner. Our job is to help you adopt new technology without giving away what makes your business valuable.

We believe you should own your assets and the business intelligence they provide for you. Mid-market businesses simply cannot afford to give away proprietary information for others to copy.

If you’d like to discuss any of these issues, schedule a call and ask for me.