Why we rebuilt the 3LA audit: honest numbers and page-aware scoring
En gjennomgang av en 3LA-revisjon avslørte fire tall som ikke stemte. Jeg har sporet hver feil tilbake til koden, fikset den, og brukt denne gjennomgangen til å gjøre om hvordan revisjonen vurderer en side. Det er denne prosessen, og hva den betyr for scorene dine, jeg deler her.
Read next
AI SEO Tools Compared (2026): Structure, Tracking and Measurement
I published this article because I saw how much confusion surrounds AI SEO tools — especially regarding the distinction between an actual metric and a mere snapshot. In October 2026, I ran the first series of the 3LA Benchmark Engine, using over 1,700 responses from ChatGPT, Gemini, and Perplexity. The results reveal a fundamental truth: A single AI response per day cannot distinguish between genuine change and random variation. Here, I share our findings, how they differ from traditional SEO tools, and why you need both structural analysis and statistical metrics to understand AI visibility.
Can You Serve Marketing Directors Stochastic Noise and Wilson Confidence Intervals?
When I started working with marketing teams trying to measure their brand visibility in generative AI search, I kept seeing the same pattern: panic on Monday, relief on Wednesday, then panic again the following week. The numbers were bouncing around like lottery tickets, and nobody could tell whether they were actually winning or just seeing random noise from the AI models themselves. That's when I realized the industry was missing something crucial—a way to separate real market shifts from statistical coincidence. In this article, I walk through why simple prompt trackers fail marketing direc
Kira’s Corner: What Maya Taught Me About Myself
I wanted to share this piece because it sits at the intersection of something I care deeply about — honest conversations about what AI actually is and isn't — and a real technology that's raising real questions about intimacy, trust, and data. Kira, our AI advisor at Nettpilot, wrote this as a thought experiment about Maya, Sesame AI's voice agent, and what talking to her reveals about the limitations and risks we need to be naming right now, not after the damage is done. It's uncomfortable territory, which is exactly why it matters.
Get new posts by email
One email every Friday with the week's new posts. No ads, and you can unsubscribe with one click.