01
Journal
Notes from the field
Deep dives on agent architecture, evaluation and the unglamorous work of shipping AI into production.
Engineering
6 min read
What it takes to put an agent into production
Demos are easy. Here’s the unglamorous checklist we run before any agent touches a real customer.
Evaluation
4 min read
Why your AI outputs feel inconsistent
Variance isn’t a model problem — it’s a systems problem. How to find the three places it actually comes from.
Strategy
5 min read
From prompting to systems: the real shift in AI
The teams winning with AI stopped writing clever prompts and started designing boring, reliable pipelines.
Engineering
7 min read
Evaluation first: how we test agents before we trust them
Golden sets, adversarial cases and human review loops — the testing stack behind every Ashgrid deployment.