AI On, In and With Solace Platform

Ask ten users what it means to have AI in a product and you will get ten answers. For some it is a chat box. For others it is an agent framework. For others an MCP server. Well, when you deliver an enterprise platform in 2026, it means a broad set of requirements.

We have grouped those capabilities into three categories based on the problems they help you solve:

• Building and running agents on Solace Platform
• Using AI assistance that’s built in to Solace Platform
• Letting agents work with Solace Platform

With this post I’ll explain what we’re delivering in each of those areas, and tease what comes next.

Build and Run Agents on Solace Platform

The first way we’ve integrated AI into Solace Platform is one many people think of first: You want to build agents, run them in production, and trust them with real business data.

That is what Solace Agent Mesh is for. It is an agent development and runtime platform, built to deploy and manage agentic AI in the enterprise. We took an engineered approach to the parts that tend to go wrong once an agent leaves the demo:

• Agent development: an agent builder, so teams are not assembling everything from scratch.
• Agent self-improvement: observability and evals, so you can see how agents behave and make them better.
• Security and governance: secure entry points and session management.
• Production ROI: a runtime designed to be operated, not just prototyped.

You are not tied to one model. Agent Mesh works with any LLM or SLM including decision models, text models and image models. It also connects to your enterprise data sources, which is where agents become useful. An agent that cannot reach your real, current data is a clever guess generator.

Put simply, Agent Mesh gives you what you need to build and deploy intelligent, real-time, data-driven agents.

Get Help from AI that’s Built in to Solace Platform

The second way we’ve integrated AI into Solace platform is for the people who run the platform every day.

We’re working toward letting you ask a question in the Solace Cloud Console and act on the answer without going anywhere else. You do not have to leave the console to dig through documentation, copy a snippet into another tool, and come back.

The same idea applies to development. AI-powered tools help you follow best practices and shorten the time it takes to build. If you read our January update on the AI Mapping Assistant in Event Portal, you have already seen an example. You give it a source payload and a target payload, and it proposes field mappings you can accept, refine and deploy. You start with a suggestion instead of a blank canvas.

What this means for you:

• Fewer trips between the console, the docs and your editor
• Best practices built into the workflow, not left to memory
• Less time on routine configuration and more on the work that is specific to your business

Let Agents Work With Solace Platform

The third way we’ve added AI functionality to Solace Platform comes at things from the other direction — instead of you using AI to operate Solace, your agents operate Solace themselves.

MCP servers give agents the tools to query and operate the platform from the tools your teams already use, and AI Skills teach coding agents how to build on Solace correctly. If your developers work in an AI coding agent, that agent can work with the platform without leaving the editor. An MCP server hands an agent the tools. Skills teach it how to use them properly. You need both, because tools without guidance are how an agent ends up calling the right function the wrong way.

This leads us to how the AI world is beginning to think of applications as headless. Not every team wants a console. AI-native teams often prefer to work from a terminal, a pipeline or an agent. However, many organizations also want to work hybrid with AI tools, so the goal becomes meeting the builders where they are. A method to build and deploy with LLMs and a method to check on the work, manage in production, and work collaboratively with UI and CLI.

• Solace Broker MCP Server is now generally available, with 47 tools for monitoring, configuring and troubleshooting Solace Cloud, software and appliance brokers. Write tools are off by default, tool access is controlled through OAuth 2.1/OIDC claims, and destructive operations are audited.
• Solace Event Portal MCP Server connects AI assistants to let them design, catalog, explore, extend and update event-driven architectures (applications, events, schemas and more) through natural language.

Finally, the capabilities you use in the Solace Cloud Console are becoming available through other doors that are open to agentic interaction:

• Management APIs and CLIs for operating and configuring the platform
• Solace Platform MCP servers for agents
• Solace Skills to teach those agents the right way to use them
• Frameworks, dev kits and specs for building custom apps, custom micro-integrations, data pipelines and Agent Mesh workflows

Under all of it sit the same Event Portal, Event Broker, Agent Mesh and Micro-Integration runtime you already know and love. Headless, GUI or CLI, the platform does the same job. Your team picks the way in.

Catch a demo on our what’s new page:

• AI Skills: Your AI Coding Agent Can Now Build on Solace Without Guessing | Solace
• Solace Broker MCP Server: Focused Tools for AI-Assisted Broker Operations | Solace

What’s Next for AI and Solace Platform

Some of what follows is still taking shape, but here’s a quick look at some details on where we see customer needs.

• Management plane MCP. We want agents to be able to work with the control plane: service lifecycle, estate health, mesh topology, SaaS operations, and more. We’ll pair them with skills so agents know how to use them well.
• Where the servers run. You should be able to run an MCP server yourself or have us run it for you, and so we are looking at scenarios that help organizations build and deploy with AI.
• Docs and support. Adding a server over our documentation and support content will make it easier for builders to find answers from inside their own tools. Every docs page is already available as Markdown, and we are looking at curated llms.txt indexes point AI tools to the right pages.

We are also exploring several other ideas:

• Schema Registry MCP for schema discovery and governance by agents
• Datadog MCP for Insights customers, making it easy to connect Datadog’s MCP server to their Solace monitoring data
• Chat in-console: scoping the key tasks where AI assistance in the Solace management plane helps most.
• AI capabilities in the new Solace Studio, starting with help creating runtimes and micro-integrations

Why Our Incorporation of AI into Solace Platform Matters

Enterprises have moved quickly from their first LLM to retrieval-augmented generation to agents, and many are somewhere in the middle of that path. The challenge now is not getting an agent to say something plausible. It is giving agents access to real business data and events in a way that is secure, governed and affordable to run.

That is the thread through all three parts of this post. Whether you are building agents on Solace, getting help in the console, or letting your own agents work with the platform, the aim is the same: put trustworthy AI to work on your real business data and events.

Next steps

• Try it: Explore Solace Agent Mesh and the AI tools in the Solace Cloud Console.
• Connect your agents: Look at the MCP servers and Skills for your coding agent.
• Tell us what you need: if you have views on the management plane MCP or the headless direction, reach out to your Solace account team.