Practical thinking on AI for business
No hype. No jargon. Just what works, what doesn't, and what GCC consumer brands should actually be thinking about.
AI Won't Fix a Company That Can't Ship
Buying AI tools before fixing your execution problems doesn't accelerate your business — it accelerates your dysfunction. Here's what to fix first.
The AI Supply Chain Has Failure Modes Your Dev Team Never Planned For
AI systems inherit every fragility of traditional software supply chains — then add a dozen new ones. Here's what solo operators and SMBs need to understand before they're blindsided.
The 98% Problem: Why AI Agents Fail Before They Start
Getting an AI agent to work once is easy. Getting it to work reliably — every time, on real tasks — is where most solo operators quietly give up. That gap has a name: harness engineering.
Plaith webhook test
If you see this on the receiving side, the wiring is good.
Where AI work lives
AI work in most consumer brand operations has nowhere to live. Agencies build it and move on. Freelancers build it and leave. In-house teams build it and get redirected. Six months later, the institutional knowledge has dispersed.
Another ai post
12 senteces about ai here
AI as infrastructure, not feature
Most companies are running AI as a feature — a tool individual users adopt. The companies winning are running AI as infrastructure — embedded in operations, used by default, invisible by design.
Operational AI is not software. Stop procuring it like it is.
The procurement process was built for software — fixed scope, one-time licence, clear deliverable. Operational AI is a working system that has to fit a business and evolve with it. Buying it through a software lens is the structural mismatch most companies haven't named yet.
Operational AI is not software. Stop procuring it like it is.
The procurement process was built for software — fixed scope, one-time licence, clear deliverable. Operational AI is a working system that has to fit a business and evolve with it. Buying it through a software lens is the structural mismatch most companies haven't named yet.
Operational AI is in 2005
Web development was once stuck in a phase where the technology worked but the products that operated it at scale hadn't been built yet. That phase was 2005. Operational AI is in that phase right now.
Want to see what could change in your operations?
A first conversation is free, takes 45 minutes, and has no agenda beyond understanding your operations.