Probably because I have been spending a lot of time in various hospitals over the last month or so, this article from the ever-reliable Microsoft Research team, really caught my attention.
In collaboration with Drexel University and the Broad Institute, they’ve developed an AI assistant aimed at helping genetic professionals diagnose rare diseases. It’s not flashy, it doesn’t come with a ribbon interface, and it won’t be appearing in Teams any time soon.
What it does do is quietly useful. The assistant reads whole genome sequencing data, flags cases that might need a second look, and pulls together gene and variant information from scientific literature. It’s not trying to be clever, it’s just trying to be helpful — which, frankly, is a refreshing change.
Now, if I were rolling this out in a real-world setting, I’d be asking some typically blunt questions. How does it integrate with existing clinical systems? Can it explain itself when challenged? And who’s accountable when it gets something wrong? These aren’t edge cases, they’re the reality of deploying AI in healthcare.
There’s also a deeper concern here. A recent study published in The Lancet Gastroenterology & Hepatology looked at how clinicians interact with AI when diagnosing gastrointestinal conditions. The researchers found that when AI support was removed, diagnostic accuracy dropped, not dramatically, but noticeably. It wasn’t that the clinicians had forgotten how to do their jobs, but rather that they’d come to rely on the AI’s filtering and triage.
A slightly different use case, I know, but this raises some uncomfortable questions. Are we training clinicians to think better, or just to think with AI? And what happens when the AI is wrong, or unavailable, or simply not trusted?
Still, the Microsoft Research assistant is a solid example of what the cloud can do when it’s pointed at something meaningful. Not the sort of thing that gets airtime at Ignite (probably), but the kind of work that actually matters..
👉 https://www.microsoft.com/en-us/research/blog/using-ai-to-assist-in-rare-disease-diagnosis/
What’s New in the Microsoft Cloud This Week
While the AI assistant for rare disease diagnosis is for me the most, interesting news this week, there are a couple of other updates, that I think are worth sharing this week, including.
Agent2Agent Protocol for Multi-Agent Apps
Microsoft has introduced something called the Agent2Agent protocol, which is essentially a way for AI agents to talk to each other without descending into chaos. It’s built to work with Semantic Kernel and LangChain, which means it’s aimed at developers who are already knee-deep in prompt engineering and orchestration frameworks.
Whether this becomes the lingua franca of multi-agent systems or just another protocol that gets quietly deprecated in two years remains to be seen.
👉 https://azure.microsoft.com/blog/azure-ai-making-ai-real-for-business/
SPFx Roadmap Update – September 2025
SharePoint Framework is getting a monthly roadmap series. Yes, really. Microsoft is promising webpack-based toolchains, open-sourced Yeoman templates, and something called “AI extensibility” for intelligent portals.
It’s a decent move, especially if you’re still building intranet sites and wondering when SPFx will catch up with the rest of the modern web.
👉 https://devblogs.microsoft.com/microsoft365dev/sharepoint-framework-spfx-roadmap-update-september-2025/
This week on LinkedIn
Unsurprisingly, this week’s updates seem very much focused (at least from my UK based colleagues) on the recent cyber attack on Jaguar Land Rover.
Azhar provides a deep dive into what we know so far and offers some valuable suggestions as to what organisations can do to mitigate the risks of falling victim to a similar attack.
and Tim offers a more business focused view in his post and link to an FT article, https://www.linkedin.com/posts/timtaylorttt_the-suspected-weak-link-in-the-jaguar-land-activity-7377322828406091776-9pmy?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAGPt4EBUQnHuhRKZiTHxz8hwxfIV_l3JTM
To be brutally honest, I think Peggy, meant to tag a different Richard at IBM (this may just be my Autism playing up though 😊). But an interesting post and podcast around cloud modernisation, and how IBM, Hashicorp and Microsoft can help organisations accelerate their cloud journey, whilst reducing complexity and increasing security.
https://www.linkedin.com/feed/update/urn:li:activity:7377688465892646912/
Am I allowed to self-publicise on my own newsletter, not sure, but please find a link to a recent article I published adding my own views to the recent Microsoft Migrate and Modernise Summit.
Cloudy with A Chance of Insights EP20
Continuing with my constant (well, not really constant…) attempt to find a good way of promoting the podcast on LinkedIn, I have created a company page to provide a single place on LinkedIn for all content, information and updates on our Pod.
Please follow to receive said updates: https://www.linkedin.com/company/108923913/admin/dashboard/
In our latest episode of Cloudy with a Chance of Insights, Cyrus and I dive into a whirlwind of updates and insights from across the Microsoft ecosystem and beyond. From critical security patches to strategic government partnerships, this episode is packed with analysis, context, and a few predictions for the future of IT careers in the age of AI.
YouTube: https://youtu.be/wkK89CkSOEU
Spotify: https://open.spotify.com/episode/63mNgiyP6oZx6A5IhWLvM8?si=2ba17c013f1442fe
Apple: https://podcasts.apple.com/us/podcast/patch-tuesday-and-smb-entra-app-management-policies/id1783369178?i=1000727821722
