Writing

The agent that nearly paid the wrong supplier

This autumn, four personal AI agents launched in seven weeks: OpenAI's dots, Meta's Muse, xAI's Grok Bot and Instinct. Each gets its own computer in the cloud, signs into your apps and keeps workin...

The interesting part of OpenAI's dots is where they put the rules

OpenAI launched dots at DevDay on 29 September. They're always-on agents, and each one gets its own cloud computer. Most coverage will lead with the cartoon avatar. I think the safety post that…

OpenAI just gave away the part that actually matters

On 19 August 2026 OpenAI open-sourced the Codex harness. The announcement contains one sentence, without a heading, that settles two years of arguing about which model is best.

Every box in the Salesforce AI stack, and why it is there

A walk through the whole architecture in plain English, so you can follow a single question from the moment someone types it to the moment an answer comes back — and explain it to somebody else afterw

Every sentence Claude writes now carries a mark you cannot see

What Anthropic switched on this month, why the law made it happen, and the one exemption that decides what it means for your business.

Anyone can generate an architecture now. That was never the hard part.

What a weekend-sized tool tells you about where the value in system design has moved, and why the expensive half is the part nobody automated.

Copilot Studio stopped asking you to write the plan

On 3 August 2026 Microsoft shipped a new engine underneath Copilot Studio. These are the stories of what it changes, what it quietly costs you, and what to do on Monday if you already have agents...

You bought Copilot. Congratulations, you're now an AI provider.

What Article 50 actually means for a company running Microsoft 365 Copilot — internally, externally, and in the agents your staff quietly built last quarter.

Your Brain Can't Tell the Difference

These are stories about one strange, hopeful fact: the mind you rehearse in private becomes the life you live in public.

The coin slot the web forgot: how AI agents are learning to pay

You don't own the intelligence you rent

In three months the ground shifted: the labs build the models, the hyperscalers rent you the whole stack. What that means for UK sovereignty, resilience and control.

The 9 AI security controls most companies quietly ignore

These are not warnings from a vendor trying to sell you something. They are nine small stories about the moment an AI system did exactly what it was told — and that turned out to be the problem.

The agent you build today won't survive the year — so build the kitchen, not the dish

These are stories about the gap between a dazzling demo and a number on a finance report — and what it actually takes to close it.

Two companies bought the same AI. A year later, only one had changed.

These are stories about a quiet split happening inside companies right now — between the ones who treat AI as a tool, and the ones who treat it as a new way to work.

Microsoft just quietly stopped calling it a Copilot

These are the stories from two weeks in which Microsoft changed the metaphor it had built its whole AI strategy on — and you can read its entire roadmap in that one swapped word.

The Machine Learned to Hunt Its Own Wounds

An AI can now read a codebase, find the bug nobody saw for twenty years, and hand you the patch. Here is what that means — and why I am both thrilled and careful.

The model is the easy part

Everyone is watching the model. New benchmark, new release, new leaderboard — the timeline lights up every few weeks. We argue about which one is smartest this month.

How to implement Microsoft 365 Copilot in your company — and why most organisations get it wrong

Six stories from the front lines of enterprise AI deployment, covering permissions, sensitivity labels, DLP, identity, monitoring, and phased rollout strategy.

How to implement Microsoft Copilot Studio in your company — the technology, the organisation, and the governance loop that holds it all together

Eight stories from organisations that got it right — and two from organisations that didn't. Every step of the way, from the first person asking 'can we build an AI agent?' to the kill switch that sto

How I gave Claude Code a memory, the same way Hermes does

Hermes is an agent that keeps notes on its own work. After each task, it reads its memory files, decides what's worth saving, and writes the answer back so the next session knows a little more than...

The EU AI Act: What It Actually Requires, and Why Most Teams Are Behind

The EU AI Act is not a future concern. Parts of it are already law. This is what it means for anyone building or deploying AI that touches European users.

Eight Stories That Will Rewire Your Mind

Dispenza's most powerful ideas — not as theory, but as stories you'll actually remember. From emotional addiction to the quantum field, told the way the brain learns best.

How Dunaway Built an AI Agent That Answers Regulatory Questions Live, in the Meeting

A Texas engineering firm saved 10,000 hours a year by replacing manual document searches with a Copilot Studio agent. Here's exactly what they built, how the architecture works, and why their engineers now answer client questions they used to defer.

MCP Changes the Game, Power Fx Gets Serious, and the Governance Gap Nobody Talks About

Copilot Studio now speaks MCP, Power Fx is becoming the computational backbone of enterprise agents, and the governance patterns you need before going to production.

Microsoft's Blueprint for Enterprise AI: What the Unified Data Platform Actually Means

Microsoft's Cloud Adoption Framework now includes a complete reference architecture for AI-ready data. Here's what every layer does, why the non-technical decisions come first, and what this means for teams building production AI.

Enterprise Agents Are Growing Up: Testing, Security, and Human Oversight

Computer-using agents now ship with secure credential handling and monitoring. MCP opens agents to human-in-the-loop oversight. And the testing gap is finally getting frameworks. This week's sources show Copilot Studio moving from prototype to production discipline.

The Threats Your Copilot Deployment Didn't Account For

Malicious browser extensions are harvesting enterprise AI chat histories at scale, threat actors are weaponising AI tools, and sensitivity labels just closed a gap most M365 environments didn't know they had. Plus: why testing your Copilot Studio agents before production isn't optional.

Before you turn on Copilot, you have a permissions problem

The security work that matters for a Microsoft 365 Copilot rollout happens before the feature is enabled — not after. Data oversharing, shadow AI already in use, misconfigured agents, and DLP policies that don't actually work: here's what practitioners are documenting.

Prompt Injection Is No Longer a Research Problem

Agentic browsers are being exploited in the wild, vendors are shipping lockdown modes for enterprise agents, and open-source security proxies are emerging. The attack surface is maturing faster than the defences.

AI adoption doesn't stall on the technology

When enterprise AI projects fail, the blockers are almost never the model. They're the platform, the operating model, the governance, and the economic assumptions the project was built on. What forward deployed engineers actually see.

The Complete Guide to Microsoft Agentic AI: From Copilots to Autonomous Agents

Microsoft has moved from AI that answers questions to AI that takes actions. This guide explains what that actually means — how the tools work, how agents connect, what it costs, and how to deploy it without chaos. Written for business leaders and IT teams alike.

The data is never what they say it is

Enterprise data has two versions: the architecture diagram, and the reality. The gap between the two has derailed more projects than any model failure ever has.

Palantir's 12-layer agentic architecture is the most ambitious enterprise AI blueprint yet

Most enterprise AI platforms bolt an LLM onto an existing data stack and call it "agentic." Palantir's AIP takes a fundamentally different approach: a 12-layer architecture where the Ontology — not th

6 AI frameworks that actually tell you how to work with AI (not just talk about it)

Six frameworks from Microsoft, Deloitte, Gartner, MIT, PwC, and Harvard Business Review all address how organizations should actually work with AI. We traced each one back to its original source.

The rise of LLM gateways: the control layer enterprise AI was missing

Every enterprise building with large language models hits the same wall. LLM gateways solve it by sitting between your applications and AI models, enforcing rules, tracking spending, and routing traff

OpenClaw and the lobster-shaped future of AI agents

175,000 GitHub stars in under two weeks. OpenClaw represents the clearest signal yet that personal AI agents have crossed from experimental curiosity into daily-driver territory.

MLES Architecture: When Your ML Code is Just the Tip of the Iceberg

Your model training notebook works beautifully. Accuracy looks great. F1 score is solid. Then you push to production and realize the model is maybe 5% of what you actually need to build.

Forward-deployed engineers: the missing link in AI adoption

95% of enterprise AI pilots fail. Not because the models are bad — they fail because nobody bridges the gap between demo and deployment. FDE job postings grew 800%+ in 2025.

Tokenomics: the hidden economics of every AI conversation

A single question to Claude Opus costs roughly 2 cents. The same question routed to GPT-4o mini? About 0.04 cents. That's a 50x difference. Welcome to tokenomics.

Microsoft's AI Agent Decision Tree, Simplified: How to Pick the Right Tool for Your Use Case

Microsoft's Cloud Adoption Framework decision tree answers the question every enterprise architect is asking: should you use a pre-built Copilot agent, build custom in Foundry, use Copilot Studio, or skip agents entirely?

Claude Cowork: Your AI Assistant That Actually Does the Work (Complete Guide 2026)

What if you could hand off your most tedious computer tasks to a capable assistant who works independently while you grab coffee? That's exactly what Claude Cowork offers.

Claude Code for Beginners: Your Friendly Guide to AI-Powered Development

Claude Skills Comprehensive Guide 2026: Everything You Need to Know About Agent Skills

Enterprise GenAI : Implementation Blueprint

What McKinsey's Reference Architecture Really Means

Enterprise AI Agents: The 2026 Strategic Playbook

How Organizations Are Transforming Operations with Autonomous AI Systems

Power BI - AI automation

The Microsoft AI Ecosystem Evolution

The Giant Leap to Small: A Guide to Edge AI and SLMs in 2025

The Giant Leap to Small: A Guide to Edge AI and SLMs in 2025For years, the narrative around AI was "bigger is better." We watched parameters balloon from billio...

Building a Specialized Medical Assistant: A Guide to Fine-Tuning and Deploying Google's MedGemma

Building a Specialized Medical Assistant: A Guide to Fine-Tuning and Deploying Google's MedGemmaThe landscape of medical AI is shifting rapidly