Nordic Semiconductor has added AI-assisted development to its wireless IoT microcontroller, with workflows covering the full life cycle from the first prototype to a deployed fleet.
Many developers copy/paste information from LLMs trained on generic data. However, Nordic’s AI solution is specifically trained on the nRF Connect SDK documentation and nRF Cloud data and integrates with a developer’s favorite IDE. It also connects to Claude Code, Cursor, GitHub Copilot, or any other LLM at a much lower token cost thanks to the specialized model.
The company says it’s based on an implementation of the Model Context Protocol (MCP), where the Nordic MCP servers give AI assistants access to validated sources from Nordic, including SDK documentation, API references, device configurations, and the customer’s field data from nRF Cloud.
Highlights of Nordic’s AI-assisted development
- Connected to nRF Connect SDK documentation and nRF Cloud data
- Integrates with AI assistants such as Claude Code, Cursor, GitHub Copilot, or any other
- Designed to assist, not replace, developers
- Covers the full development lifecycle, from prototyping to fleet management
Nordic explains the AI agent can be especially useful to automate tedious tasks, speed up prototyping, and ease debugging, for instance, when migrating between SDK versions, for custom board bring-up, or diagnosing a crash on a deployed device.
Several video examples are provided, including AI-assisted migration, finding faulty devices in your fleet, keeping AI costs down, troubleshooting user-reported errors, DeviceTree and Kconfig generation (see video below), validating release readiness, and adding shell commands.
The important part is to be specific and review the code, since the agent can make mistakes. For example, in the video above, the AI agent added some random peripherals (a button and extra LEDs), and this had to be corrected manually. It’s also clear you have to be an engineer to use these tools, since the prompt needs to be quite specific and technical, or in other words, you can’t just “vibe code” your way to a board bring-up. I still view AI agents as interns or drunk/high senior software engineers; in either case, you can’t just let them do their own thing, and they need supervision…
AI is often seen as a threat to the livelihood of software engineers, but so far, it’s mostly a tool. While some companies have fired software engineers due to AI, I’ve seen several headlines about rehiring due to prohibitive AI costs. Some also argue that the number of software engineers may increase rather than decrease, since cheaper software, and in this case firmware, development costs may eventually lead to higher demand for software. Time will tell.
You can check the examples and learn how to get started on the Nordic’s website.
Thanks to TLS for the tip.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
