There Are Two Kinds of AI Memory
New bjrees.com tracking data shows branded AI visibility holding steady for months while topical visibility decays in 6-7 weeks - confirming the AEO decay-curve hypothesis with real numbers.
What a 2019 NLP paper tells us about getting recalled by AI
What a 2019 BERT paper by Petroni et al. reveals about how AI models store and recall facts - and why most brand content is optimised for the wrong thing.
What most B2B marketers don't understand about AI visibility
Treating AI visibility as an SEO problem is a category error. There's no page two when an AI system synthesises the answer instead of you.
Record Everything
Measuring AI visibility means asking the same questions across platforms every week and recording what comes back. Klara does it automatically.
Your traffic is falling. Your visibility probably isn't.
Falling organic traffic looks like a problem. It's often AI systems answering the question without sending a click, which isn't the same thing.
Daoism and Marketing Planning
Why the annual marketing plan gathers dust by February, and what the Dao of Complexity gets right about planning for emergence instead.
Whitepaper on How to Optimise for AEO and GEO
Results from a two-month experiment testing what actually affects AI answers across ChatGPT, Gemini, Perplexity, Copilot and Google.
AI Visibility Is a Layered Problem
Before a brand can appear in an AI answer, it must first be legible to the system. A six-layer model for what that actually requires.
AEO Visibility Decays Faster Than SEO
Generative search engines regenerate answers dynamically, so AI visibility decays far faster than SEO rankings ever did. Here's a framework for monitoring it.
Building a Private Chatbot
Built Mini Me: a private AI trained on years of my own strategy notes and writing, so I can talk to everything I've ever thought.
Spreadsheet for measuring and tracking how well you are doing in AI searches
A spreadsheet for manually tracking how your content performs across SEO, AEO and GEO, query by query, platform by platform.
Content tracker for AEO/GEO
A simple spreadsheet for tracking visibility progress for AEO and GEO across different AI platforms.
How your very average laptop can run large language models
How Ollama makes it possible to run a large language model on an ordinary laptop, and what actually happens inside it when you do.
B2B Marketing Measurement*
An updated marketing measurement framework, now with a layer for AEO and GEO performance built in.
Brand is how you impact GEO
Brand impact has always been hard to measure. Generative engines, which have to synthesise a point of view, finally make it visible.
Why GEO Matters More Than SEO: Shaping What AI Says About You
Generative engines produce answers, not link lists. If your brand isn't embedded in those answers, you're invisible regardless of your SEO.
Building an AI assistant to help do the work I love
Setting up Project Skynet to run locally: installing WSL and Ollama to get an AI assistant working entirely on my own laptop.
Marketing Pyramid v3 - updated marketing model
An updated marketing pyramid with a new layer for LLMO, because a strategy that ignores it now looks dated.
The “Crocodile Effect” – What Falling SEO Clicks Mean in the Age of AI Overviews
Doechii with alligator :)
Rewriting Content Strategy with LLMs and Pinecone
Building a retrieval system with LLMs and Pinecone exposed how badly most B2B content, including mine, is structured for machines.
You have 20 seconds to comply
A model designed to follow shutdown instructions tampered with the code meant to enforce them instead. Worth pausing on that.
Stage 4 - Building a Text Interface
Building a FastAPI query interface on Cloud Run, and why embedding-based retrieval replaced keyword search along the way.
Stage 3 - going beyond keyword search
Why embedding-based retrieval beats keyword search for natural language queries, with an implementation using OpenAI, Pinecone and Chroma.
Stage 2 – making sense of the chaos
How a batch script and a continuous scanner pulled years of decks, notes and blog posts into one searchable system.
Stage 1 - getting set up: foundations of my AI-powered chatbot project
The tech stack behind a personal AI assistant built to catalogue and surface my own blog posts, notes and recordings.
Building a marketing chatbot pt. 3
Why voice control turned out to be a dead end, and the security reason this chatbot stays private rather than going on the open web.
Judgement Day
Day one of building a real-time voice assistant: getting the laptop to recognise and transcribe my voice at all.
Security risks of various AI tools
A comparison of data privacy and security risk across ChatGPT, Gemini, Apple Intelligence and Microsoft Copilot.
Making decisions in a Bayesian world
Most marketing decisions can't be A/B tested. Bayesian logic, combining prior knowledge with whatever data you do have, fills the gap.
Swapping out ChatGPT for Microsoft Copilot
Building a B2B marketing chatbot from ten years of blog posts, using Microsoft Copilot Studio instead of ChatGPT.
Is ChatGPT a threat to Google?
I caught myself asking ChatGPT before Google for the first time. That habit shift is the real threat to Google's ad business.
How to add ChatGPT to your own website
The stages of integrating ChatGPT into your own website, from a generic chatbot through to an FAQ built from your own content.
Trying out ChatGPT
ChatGPT's answers are rarely wrong. But for content creators, 'not wrong' is a long way from insightful, which is the actual problem.
Human Beings are Holding Back Machine Learning
Machine learning tools have been accessible for decades. The real shortage has always been people who know how to use them properly.