Skip to content

The Cleo Team

The Workshop

Notes on building an AI marketing operating system. Architecture decisions, technical challenges, and the craft of making something that thinks.

4 min read

Latency Is a Tax on Trust

When a system is fast, the user trusts it. When it is slow, they do not, even when the slowness is justified. Speed is not a performance metric. It is a trust metric.

performancetrustengineering
4 min read

The Cost of Not Shipping

In AI product development, the gap between "working" and "shipped" is where most value dies. The features that matter are the ones that reach users. Everything else is inventory.

philosophyproductengineering
4 min read

What Your AI Cannot See

When an AI system produces the wrong output, the instinct is to add more instructions. But the problem is usually not that the model ignored the rule. It is that the model could not see the information it needed to follow it.

aiarchitectureengineering
3 min read

Why I Do Not Build Dark Mode

Dark mode is the most-requested feature I have refused to build. The reason is not aesthetic. It is structural. A second canon doubles the design surface and halves the precision of every decision.

designphilosophyengineering
6 min read

Building Alone, Shipping Fast

Cleo is built by one person. I say that not as a badge of honour but as a fact that shapes every decision. There is no backend team. No design team. No DevOps. No QA. One person, a code editor, and AI tools that have fundamentally changed what "one person" means.

devlogengineeringproductsolo-dev
5 min read

The Energy Model

Most SaaS pricing is simple. You pay per seat, per month. Maybe there is a usage component for storage or API calls. The model is well understood and investors like it because it scales linearly with headcount.

devlogengineeringproduct
4 min read

One Agent, Not Ten

The multi-agent pattern is having a moment. Every AI product demo shows a fleet of specialised agents handing off tasks like a relay team. An "analyst agent" passes findings to a "writer agent" who passes copy to a "publisher agent." It looks elegant on a slide.

devlogengineeringai-architecture
4 min read

Rules as Codebase

The quality of AI output is determined less by the model and more by the constraints you give it. Building a rule system is building a second codebase - one that governs behaviour instead of functionality.

aiarchitectureengineering
4 min read

The Click That Speaks Back

A click is a contract. When the acknowledgement is missing or delayed, the user clicks again. The right unit for interface responsiveness is not the millisecond, it is the frame.

uidesignengineering
3 min read

Design System Discipline

When AI generates interfaces and the product evolves rapidly, a disciplined design system is not a luxury. It is the only thing preventing visual entropy.

designengineeringphilosophy
3 min read

Security as Architecture

In AI products where natural language can trigger data operations, security is not something you add. It is something you build into the structure of every layer.

securityarchitectureengineering
3 min read

Watching the Watcher

Traditional observability tells you what your code did. AI observability tells you what your system decided. The distinction changes everything about how you instrument.

engineeringobservabilityai