An AI agent can query every database in your company and still make the wrong decision.
Not because the model is weak. Because Sales, Finance and Operations may have three different definitions of the same number.
That is an architecture problem. And a leadership problem.
Two announcements last week caught my attention. Oracle outlined an AI Data Platform built around business ontologies, semantic reasoning and governance. Zoho announced semantic and ontology layers for its analytics platform.
This is not a new direction for us at Microsoft. Since Ignite 2025, Microsoft Fabric IQ has included Ontology in public preview, providing a semantic foundation that connects business entities, relationships, rules and data so people and AI can reason from a shared understanding of the business.
Different platforms, same signal: connecting AI to data is not the same as connecting AI to the business.
Imagine an agent asked to prioritise the most profitable customers. Does “profit” mean gross margin, contribution margin or lifetime value? Are returns and rebates included? And who counts as the same customer across CRM, ERP and billing?
The agent might produce an elegant answer. It could still be commercially wrong.
I think enterprise semantics should be treated as shared business infrastructure, not a technical clean-up exercise.
That means accountable business owners for critical definitions, versioned data contracts, traceable lineage and policy, and tests to verify that agents apply those definitions consistently before they act.
The technology matters. But the harder decision often comes first: getting leaders to agree on what a metric means, who can change it and which interpretation takes precedence.
Before funding another AI initiative, I would pick five decisions that materially affect revenue, margin or customer experience. Trace the data and definitions behind each one. If the definitions are disputed, that is the first backlog to resolve.
We talk a great deal about making AI more intelligent.
Perhaps the higher-return move is making the enterprise less ambiguous.
Who owns the definition of business truth in your organisation?
#EnterpriseAI #DataArchitecture #TechnologyLeadership
2
127
allerhed retweeted
Today at Universe we talked about the changes we are making to git to support collaboration at machine speed. The new Git architecture will support 7,200 writes / sec!!! Coming early next year.
allerhed retweeted
We now have state-of-the-art pre-trained base models in the 3T-parameter weight class, built entirely in-house at MAI.
Post-training results coming soon -- stay tuned!
Lots of details in our technical blog post.
microsoft.ai/news/frontier-p…
Tomorrow: GitHub Copilot Day 🎉
New releases, real workflows, and live coding, directly from the team who builds it and the developers who put it to work.
xitter.gq/i/broadcasts/1RKjpbkXE…
allerhed retweeted
We just launched Xirp, a vendor-neutral agentic development environment. One place to manage agent sessions across @ClaudeDevs, @GeminiApp CLI, and @OpenAI Codex. 1,300+ @Spotify engineers already use it. Now it's available for you to try. Learn more at xirp.spotify.com.
allerhed retweeted
Some more detail on the ROIC Intelligence App I built yesterday and mentioned on today's earnings call.
I took the PDF that Brian Nowak at Morgan Stanley put together for Hyperscale ROIC this week and used Copilot code (coming in our new superapp) with a single prompt + skill (/drill-me) to create the plan, then used autopilot in auto to create the full app (with history, lookups, scenarios, what-ifs, etc). And /rubber-duck to test.
And the best part is that all the artifacts are in my enterprise environment. My app is in Copilot, my code is in GitHub Enterprise; all my data pipelines/lake/semantic models are in Fabric. And everything is under Agent 365 IT/Sec/FinOps control!
So this is not about Tokenmaxxing or vibe coding. Every step of the way the rails are engineered to create value, making everything a long-term reusable asset, with governance/security, and cost controls.
This is the full system to drive business value. Disclosures: This is all pulled from public sources, and for illustrative purposes only...not financial advice! :)
Here is the app and architecture...
allerhed retweeted
We just reported our FY26 Q4 earnings.
🟢 A strong quarter with $90 billion in revenue
🟢 Azure and other cloud services revenue grew 43% driven by strong demand for services across the platform
🟢 Microsoft 365 Copilot reached more than 30 million paid seats
“We are advancing the frontier on the cost-to-outcome curve, ensuring every customer can turn tokens into business results. This year, Azure revenue surpassed $100 billion for the first time, and Microsoft 365 Copilot reached over 30 million seats, reflecting the confidence customers are placing in us to power their AI transformation.” - @SatyaNadella
Read the full release here: msft.it/6016a610w
allerhed retweeted
Today, we are announcing a series of updates that give customers frontier-grade security at half the cost.
MAI-Cyber-1-Flash is our first cybersecurity model, built ground up to find the most challenging vulnerabilities in complex code bases. When combined with MDASH, it delivers world-class performance at 50 percent of the cost of leading models.
We are bringing this capability to market through Project Perception, a complete agentic security offering grounded in real-world signals and security workflows. Teams of specialized agents work together to simulate attacks, detect and triage/investigate, and fix and remediate.
This is the benefit of building the harness, context/signals, and action space separate from one model family. By combining specialized models and data with the right agents, tools, security context, and harness, we can advance the frontier of cost to outcome.
allerhed retweeted
The future of the firm is a learning loop in which human capital and token capital compound.
With our new Frontier Co., our ambition is to help every enterprise build its own AI capability, and to help create a frontier ecosystem where every organization can turn its knowledge, workflows, and judgment into its own AI systems that continuously improve. blogs.microsoft.com/blog/202…
allerhed retweeted
Huge news. Anthropic models now GA in Foundry.
Claude in Microsoft Foundry is now generally available, hosted on Azure. Prompt caching and extended thinking are supported today, with more capabilities on the way.
Read more: msft.it/6018vsDhM
allerhed retweeted
Today we’re bringing skills to Copilot for Excel, giving teams a new way to scale their expertise across every workbook.
allerhed retweeted
I'm hiring product and AI engineers to work on the GitHub Copilot App.
If you're the right person to invent future of work, reach out to me with some proof.
If you work at GitHub/Microsoft and this sounds appealing to you, reach out to me internally with similar artifacts.
allerhed retweeted
Introducing Surface Laptop Ultra.
Built for world makers. Designed for what's next.
The most powerful Surface laptop ever. Coming Fall 2026.
Sign up to learn more: msft.it/6019vw79T
allerhed retweeted
If you've been wondering what I've been up to, the team and I have been cooking up something new.
A new agent-native development environment deeply integrated with the GitHub graph. Not just for writing code, but all of the meta-work as well.
allerhed retweeted
Copilot Studio just got a major upgrade with a new drag-and-drop workflow designer (Preview)🔥🔥🔥
You can now build complex automation and agentic reasoning on a single canvas⚙️🧑💻
🔗 New workflow designer guide: learn.microsoft.com/en-us/mi…
#CopilotStudio #PowerPlatform #AI
allerhed retweeted
Copilot Studio is getting DANGEROUS.
New flow designer showing in Preview.
Looks a lot closer to Make and n8n.
Classify. Route. Call agents. Write back.
Microsoft is coming for the automation canvas.
allerhed retweeted
New in Copilot Cowork: mobile, skills, and plugins.
Now available on iOS and Android, so you can delegate work from your phone, pick it back up on your desktop, and keep tasks moving without breaking flow.
And with new connectors, Cowork can operate across business systems and data.
allerhed retweeted
Copilot Cowork SURVIVAL TIP.
Tell Copilot Cowork to create a requirements.md file at the start of your session.
Not later.
At the start.
Have it track:
decisions,
requirements,
open items,
conversation context,
audit notes,
files created.
Then if the session glitches or you need a fresh chat, open the Output folder, copy the OneDrive link, paste it into the new session, and tell Cowork:
Use the files from our previous session.
This has saved me more than once.
Cowork loves creating files.
Make it create the one that keeps your work alive.
Mini blog post link in the comments👇
🤖 Made with AI





















