Adam Listek
Devolutions Remote Desktop Manager 2025.1 introduced AI integrations built on Microsoft Semantic Kernel and exposed throughout RDM to assist with scripting, comment and translation generation, and a built-in AI assistant chat integration. Check out our previous spotlight article and set up RDM quickly with the documentation and Devolutions Academy course.
- Model choice is yours: You can choose from OpenAI, Google Gemini, DeepSeek, Mistral, local Ollama models, and more, allowing you to align with internal compliance rules and use the best available model.
- AI-assisted script generator: Script generation helps IT professionals streamline task automation. With the ability to generate code directly within a script entry type, you can take advantage of the latest AI models and work faster and more efficiently within RDM.
- Translations and proofreading assistance: In several text fields, RDM directly integrates your chosen AI model to translate text to different languages and perform operations such as condensing or proofreading your comments.
- Setting up is straightforward: Choose the model, add your API key or URL, and use a built-in AI chat directly from the main menu or within entry fields. You can also create AI-specific entry types to use different models and integrations.
The bottom line is that an AI assistant is no longer a novelty; it saves valuable time on routine tasks for teams that live in RDM all day.
Help us focus on what matters most to you
We want to hear from you! There are many directions to take AI integrations in RDM and the broader Devolutions product ecosystem, but the focus is on what’s most important for IT professionals.
Which AI-powered capability would you like to see next in Remote Desktop Manager?
- Generative bulk-creation
- AI integration with terminal entries
- An RDM/DVLS/Hub policy assistant
- AI-driven troubleshooting
- Something else?
A Sneak Peek: AI + RDM MCP
You may have heard about MCP (model context protocol) as it pertains to AI, but what is MCP, and how can it help you in RDM? MCP is a JSON-RPC protocol that describes how an AI model can interact with outside tools and systems. In this way, AI models don’t rely solely on documentation to guess and infer, but use a standard format to quickly find the best tool and command for a given request.
What does this mean for RDM? Well, we are working on an MCP for RDM that, when paired with an AI model of choice, you can efficiently instruct AI to perform actions in RDM.
Automatically generate folder and entry structures: “Create unique folders for five users containing protocol-specific sub-folders.”
Populate entries from files: “Using the provided CSV, create entries for each server and populate with asset information.”
Connect with other tools: “Pull a report of all servers from the CMDB and update the associated asset information in RDM.”
These initiatives build directly on the existing AI integrations in RDM. With MCP, RDM moves from suggesting to direct action, saving you time to focus on the complex problems.
Why MCP in RDM matters
- Time-to-value: Set up a new environment in minutes instead of spending hours on manual entry.
- Fewer mistakes: AI generates consistent names, paths, and tags.
- Audit-ready: Operations are logged through RDM’s standard activity trail.
As MCP is rapidly evolving, users of the RDM MCP server need to consider security. It is strongly recommended that they leverage a local self-hosted LLM option when using the MCP server.
We are actively developing our MCP integration, but your feedback on the poll above will help us prioritize which capabilities to focus on first.
Marc Beausejour
Matt Miller
