Notes on content trust, AI slop, and agent workflows.
Long-form posts on the design choices behind VeracityAPI, lessons from ten years at Clearscope, and the operational patterns I see in agent content systems. New posts roughly monthly.
SynthID and VeracityAPI: the layered verification stack content teams actually need
Hard provenance signals like SynthID and workflow-risk APIs like VeracityAPI answer different questions in the same pipeline. Why most teams need both layers, where each one carries weight, and what to do when the strong signal isn't available.
What 10 years at Clearscope taught me about AI slop
The lesson that shaped VeracityAPI's design: specificity, not authorship, is the signal that predicts whether content earns trust.
The most expensive bug in agent content workflows
The unbounded revise loop. Why it happens, why your agent will burn through hundreds of dollars on a single draft if you let it, and the four-line fix.
Why the AI-detection category is splitting
The category that started as 'is this written by AI?' is becoming two categories. Where each one is heading, who they're built for, and why a single product can't do both jobs well.
Why we don't publish competitor benchmark numbers (yet)
The most common question I get is 'how does VeracityAPI compare to GPTZero on accuracy?' The honest answer is that I don't have a number I'm willing to publish — and here's the system we're building to produce one.
Benchmarking AI detectors on the routing decision production teams actually need
Why VeracityAPI will report binary flagging metrics alongside workflow-routing F1, with caveats and reproducibility gates.
VeracityAPI is not an AI detector. It is a routing linter for AI outputs.
The product wedge: action plus evidence for agent workflows, not forensic authorship proof.