Skip to main content

Pragmatic AI for your business

Honest analysis of AI applied to real operations. No hype, just judgment: what you control, what you can prove, and what to build that your competitors can't.

Agent Autonomy Is a Liability, Not a Feature You Buy
Business Strategy
· 7 min read

Agent Autonomy Is a Liability, Not a Feature You Buy

Cognition raised $1B at a $26B valuation for an autonomous coding agent. In production, autonomy is the first thing that breaks. How much should you give it?

AI agents autonomous agents software architecture
Your marketing asks AI what to cut. Ask this instead.
AI in Marketing
· 7 min read

Your marketing asks AI what to cut. Ask this instead.

Box created 13 new AI roles, one to market to industries it could not staff before. The question is not what AI lets you cut, but what it makes possible.

AI marketing marketing strategy Box
Your agent invoiced $0.00. The logs never saw it.
Business Strategy
· 8 min read

Your agent invoiced $0.00. The logs never saw it.

A sales agent wrote a flawless $0.00 invoice for a 14-seat deal. The pricing API was correct. Every log was clean. Only the full execution trace explained it.

AI agents AI observability AI governance
The pyramid before the agent: you almost never need one
Business Strategy
· 7 min read

The pyramid before the agent: you almost never need one

The 'AI consultant' title expires. What stays is a discipline: start with a deterministic workflow and climb to an agent only when the problem demands it.

automation pyramid AI agents deterministic workflows
Kirkland Isn't Building a Model. It's Building a Layer.
Business Strategy
· 7 min read

Kirkland Isn't Building a Model. It's Building a Layer.

Kirkland & Ellis is putting $500M into AI and half of LinkedIn misread it: they aren't training a model, they're building the layer that sits on top of it.

Kirkland & Ellis AI moat model vs layer
What Your Team Types Into AI Is Now Legal Evidence
Business Strategy
· 8 min read

What Your Team Types Into AI Is Now Legal Evidence

Two 2026 court rulings confirmed it: what your team types into ChatGPT or Claude is discoverable evidence. The problem isn't AI. It's having no policy.

AI legal evidence AI governance shadow AI
Claude runs a thousand agents. Judgment doesn't parallelize.
Business Strategy
· 8 min read

Claude runs a thousand agents. Judgment doesn't parallelize.

Anthropic shipped Opus 4.8 with Dynamic Workflows: hundreds of parallel subagents that check each other. The one thing that doesn't parallelize is judgment.

Dynamic Workflows parallel subagents Claude Opus 4.8
He Sells AI Agents. He Told His Team to Stop Using Them
Business Strategy
· 6 min read

He Sells AI Agents. He Told His Team to Stop Using Them

Gumloop's founder told his team to stop automating everything with AI. The expensive failure of agentic AI isn't cost, it's the slop that loses customers.

AI slop AI automation AI agents
Your marketing AI bill is up 13x. Who's measuring?
AI in Marketing
· 9 min read

Your marketing AI bill is up 13x. Who's measuring?

Business AI token spend is up 13x since January 2025. Marketing is the second-heaviest user and the worst at measuring it. Build a FinOps layer first.

marketing FinOps marketing automation cost per MQL

Stay up to date with AI and digital transformation

Join our weekly newsletter with practical analysis and tips

We respect your privacy. Unsubscribe anytime.

Frequently Asked Questions about Our Blog