If you’re using Tanzu Platform, you have pretty much everything you need to start getting AI into your apps. That’s the main point I wanted us to get across in our Tanzu annual update:
Looking at surveys, it feels like most large organizations are still figuring out their AI strategy through pockets of messing around - which is totally fine! But, they need to start scaling it out to more and more - even hundreds of apps. AI needs to become “boring” as the Kubernetes Kids used to say.
How do you do that? Well, hey now, if you’re already using Tanzu Platform, it already contains the essential building blocks for scaling AI across your app portfolio.
Our overview walks through the three things that make using Tanzu Platform for enterprise AI straightforward:
Model Access Management: Tanzu’s AI service provides governance and controlled access to both public and private models–creating a single connection point for all your applications. Once you have a broker/middle-person/chokepoint in place, you can start applying all the thrilling controls, security, performance management, and other “day two AI operations” you need for wide use.
Java Integration Through Spring AI: Rather than forcing your teams to learn Python, Spring AI lets your Java developers enhance existing applications with AI capabilities using familiar patterns and tools. Each year we see in surveys a huge number of enterprise apps are done in Java, and huge number of those use the Spring Framework. Java is the language of enterprise apps, it’s what developers know. Why ditch all that and slow down to completely re-skill your people?
Python is perfectly fine, too: From what I can tell, most of the existing AI apps are in python. Which is fine, good for them! The Tanzu Platform is layered like any responsible PaaS and fully supports python as well. In fact, having the Tanzu Platform in place makes it even better because you can start to pull in all the different DIY stacks and accidental platforms you probably have now. Keep the python, ditch the chaos.)
Enterprise Context Connection: The platform now supports the Model Context Protocol (MCP), allowing your AI applications to securely access your organization’s proprietary data and systems. I’ve done doing a lot of first hand work with MCP to see if it’s a bunch of bullshit or real. There’s still a lot of work to be done, mostly for security, but they’re close and the standard is thoughtful. I think it’ll be a good framework for agentic AI and just enterprise AI in general.
What I found most valuable during the demo portion was seeing how platform engineers and developers collaborate through this approach. The platform team handles model provisioning, security, and compliance, while application developers focus on implementing business logic and features.
As Adib explained in his segment, success with AI depends on rapid iteration through ideas. The platform enables this by abstracting away infrastructure complexity and standardizing access patterns to AI services.
The result is that your teams can implement, test, and refine AI features using existing development workflows–applying the same DevOps practices you’ve already established for conventional applications.
If you’re interested in seeing this in action, including a demonstration of building an AI assistant that progresses from basic chat functionality to secure, agentic capabilities, check out the full presentation.
My co-worker Camille has been writing a lot about our AI thinking in Tanzu. First, she explains why rapid iteration is the key to getting agentic applications out of the proof-of-concept stage and into production, where they can actually deliver value. She also took a closer look at Model Context Protocol (MCP). In another piece, she breaks down the top three ROI goals orgs should target with agentic AI: cost savings, operational efficiency, and revenue lift. And over at The New Stack, Jonathan and Mark Pollack go over what developers actually need to build AI apps that make it past the demo stage.
Next month, we’re doing a day long workshop on all of this Tanzu AI stuff. It’s May 13th, in Palo Alto California. There’ll be a lot of in-depth explanation of the above, demos, coding - seeing all the stuff! I’m the MC for it, but I’ll also go over my MCP for D&D work and findings. It’s, of course, free to attend. You should come! Register for it and I’ll see you there. Also, Cloud Foundry Day is the next day, May 14th, so you can go to two great enterprise platform conferences.
That’s it for now. The usual links and strange finds are on the way next.