Blog

Autonomous AI, in practice.

Practitioner guides on running AI employees, agent governance, cost, and where AI actually pays off for SMBs.

Field Notes · Integrations

When Your Software Has No API, Put an AI Employee on the Screen

A lot of the software small businesses run on has no API to plug into. Computer-use AI can operate it by clicking and typing like a person. That’s powerful — and easy to misuse. Here’s the honest version.

Field Notes · Human-in-the-Loop

How Many AI Employees Can One Person Actually Supervise?

You can spin up a tenth AI employee in minutes. What doesn’t scale is the one person who has to approve what they all do. Approval bandwidth — not compute — is the real ceiling.

Build Log · Owned Models

Build Log: Fine-Tuning a CFO Brain That Won't Make Up Your Numbers

We fine-tuned an open-weight model into a CFO advisory brain — and failed our own audit on the first try. Here’s the full build log: what broke, what fixed it, and what it cost.

Build Log · Owned Models

Build Log: How Our Marketing LLM Stopped Making Up Statistics

Version one of our marketing model passed its evals and shipped. Then we counted how often it invented statistics: 21 outputs out of 50. The problem wasn’t the model. It was the diet.

Owned Models · Research Notes

How We Chose a Base Model for Our Open LLMs (It's Not About the Benchmark)

Everyone asks ‘which model is smartest?’ After benchmarking, a blind bake-off, and one closed-weights dead-end, we ended up asking a different question entirely: which weights can you own, tune, and serve — on the same path?

AI Employees · Playbook

One Brain, Many Agents: Why Your SMB's AI Team Needs Shared Memory

Give each agent its own memory and you get five agents with five versions of the truth. The real unlock is one brain they all read from — and write back to.

Governance · Guide

Your AI Remembers Your Customers — Now What Does Privacy Law Say?

The moment your agents keep durable memory of customers, you’ve created a small database of personal data — and the rules that govern databases now govern your AI.

Sober Take · Playbook

Stop Your AI From Making Up Facts About Your Own Customers

An agent that invents a customer’s plan, price, or renewal date doesn’t just embarrass you — it burns trust you can’t easily rebuild. The fix isn’t a smarter model. It’s grounding.

AI Employees · Playbook

AI Agent Memory: How to Give an SMB Agent Context That Compounds

The wall every SMB hits after the demo: “my agent forgets everything.” Here’s what durable agent memory is, what to store, and what to never store.

Governance · Playbook

Human-in-the-Loop AI: Where to Put Approval Gates in Agent Workflows

Autonomy without a seatbelt is just recklessness. Here’s exactly where to put human approval in an agent workflow — keyed to how much damage an action can do.

FinOps · Guide

What Does an AI Agent Actually Cost to Run? An SMB Budget Guide (2026)

“Why did last month cost 4x?” is the fastest way to kill trust in an agent. Here’s how to model and cap what an AI agent actually costs to run.

Sober Take · 2026

Should Your SMB Let AI Answer the Phone? A Sober 2026 Take

AI can pick up the phone now. Whether it should pick up yours is a different question. An honest, no-hype read for 2026.