Inside the Intelligence Platform

Insights, frameworks, and lessons.

How intelligence compounds — and why it matters.

♟
Perspective
Everyone's an agent now
Read the May 2026 hero lines from the major marketing automation vendors back-to-back and the messaging convergence is visible in real time. The agent metaphor has won. Autonomous has replaced automated. MCP has reached the marketing layer.
May 3, 2026 9 min read
⤴
Perspective
When the model layer goes public
An OpenAI listing doesn't change the models. It changes the incentives of the people who make them. The application layer needs to plan for a model layer that has just become a public-company commodity supplier.
May 3, 2026 8 min read
⊘
Perspective
Nobody is winning enterprise AI yet
The Microsoft Copilot ROI shortfall is the canary, not the failure. Eighteen months in, there is no playbook — only honest discipline and calibrated bets.
May 2, 2026 8 min read
⓪
Operations
Day zero
What we learned shipping our first customer install. Every assumption we had about deterministic deployment got tested in eight hours, and most of them failed loudly. The product is better for it.
May 2, 2026 8 min read
▣
Governance
Regulation is the moat
For an AI startup raising in 2026, governance is no longer a cost centre. It is the most defensible thing on your cap table — earlier, cheaper, and harder to copy than any feature.
May 2, 2026 7 min read
⊙
Governance
Vibes don't comply
Article 50 is what people are talking about. It is the smallest part of the EU AI Act. Articles 9 to 15 define the regime — and they reward orientation, not speed.
May 2, 2026 9 min read
▢
Positioning
Vibes don't ship. Systems do.
Most AI rollouts produce fluent text that no-one downstream can audit. The smallest set of explicit components that makes a piece of work reviewable end-to-end is the difference.
May 1, 2026 7 min read
☰
Discipline
We migrated to ADO. Here is what actually changed.
Five days after making Azure DevOps the system of record for everything we are building, what was previously impossible is now possible.
May 1, 2026 6 min read
🔍
Case Study
We forgot our own engineers
We built ORCA to reduce cognitive load, then designed an install that pushed it straight back onto the engineer. A case study in catching our own contradiction, and the single-container pivot that fixed it.
Apr 22, 2026 6 min read
⚡
Case Study
The workforce caught a bug in its own code last night
Hours before the first customer deploy, the platform reviewed its own connector code overnight and flagged a real security bug for the Founder to fix before ship. The self-improving loop closed, for the first time, in the real world.
Apr 22, 2026 6 min read
👥
Workforce
Meet the workforce
Fifteen named colleagues. Each with a defined role, a private memory, a shared brain, and a clear line manager. Not a chatbot. Not an agent swarm. A workforce.
Apr 22, 2026 8 min read
§
Governance
Your AI governance architecture IS your Art.50 strategy
The EU AI Act's transparency deadline lands in August 2026. Most organisations are treating it as a labelling problem. It isn't — it's provenance.
Apr 18, 2026 8 min read
⚬
Architecture
If you're not the model, you're the harness
The frontier labs ship models. Everyone else ships the harness — the scaffolding that decides what an agent remembers, what it can do, and what it can touch.
Apr 15, 2026 9 min read
↻
Framework
The side that loops fastest wins
The OODA loop isn't a military relic — it's the operating system for any organisation that wants to think faster than its competitors.
Apr 9, 2026 8 min read
≡
Intelligence
Knowledge Equity — the economic footprint
Not a score. Not a badge. An economic footprint that grows every time your team captures what they've learned.
Apr 7, 2026 6 min read
⌘
Governance
Governance is code, not policy
Why constitutional AI governance — rules enforced in architecture, not handbooks — is the only model that scales.
Apr 5, 2026 7 min read
✗
Perspective
The AI Typewriter Trap
Most organisations are using AI to produce more. The ones that win will use it to know more.
Apr 3, 2026 5 min read
☂
Architecture
Why most AI knowledge systems fail GDPR
PII tokenisation, envelope encryption, and the three-layer defence that makes ORCA GDPR-ready from day one.
Apr 1, 2026 9 min read
↗
Compounding
Week 1 is the worst it will ever be
The compounding effect explained — why intelligence systems get exponentially more valuable over time.
Apr 1, 2026 6 min read
☰
Reference
ORCA Glossary
Every key term in the ORCA intelligence platform defined — practice brain, confidence classification, knowledge equity, and more.
Apr 2026 3 min read