Product Agentic Pages Prompt Monitoring Pricing Blog Book a demo
Log in Sign Up Free

Field notes from the agentic web.

Research, playbooks, and benchmarks on how AI platforms find, read, and cite the brands that show up. Written by the team building Klove.

Fresh off the desk.

Sort
Long corridor of server racks in a data centre, lit in cool blue-grey, emphasising industrial scale and geometric depth.
Essays Research 29 May 2026

How ChatGPT actually searches the web.

The interface shows one model thinking. What runs underneath is a nine-stage pipeline of classifiers, semantic rewriters, and hard fetch timeouts — and most content never survives it.

11 min read Michael Shaskey
A printed market evaluation matrix on cream paper, annotated in pencil with some cells circled in red, lit by warm directional light.
AI Search Research 14 May 2026

The AEO/GEO platform market in 2026: separating dashboards from drivers.

$300M has poured into AI search optimisation tools. Most of it bought monitoring dashboards. Here's how to tell the difference before you sign.

9 min read Michael Shaskey
Dark-mode server log showing AI crawler user-agent strings including GPTBot and PerplexityBot making requests to a website.
Playbooks Playbook 3 May 2026

How to optimise your site for AI citations.

AI referral traffic grew 796% in two years and converts higher than organic search - yet only 11% of cited domains appear across multiple platforms. Here is the step-by-step playbook for earning citations from all three.

11 min read Michael Shaskey
Split view of a browser-rendered SaaS page versus the sparse HTML shell that AI crawlers actually receive, illustrating the rendering gap between human and bot visits.
Agentic Pages Essay 2 May 2026

The two-tier web: why agentic pages are the only reliable fix for AI invisibility.

Most GEO advice stops at content. It ignores the moment the crawler arrives - and leaves nothing for it to read.

10 min read Michael Shaskey
Close-up of a network patch panel in a data centre, multiple cables connected to numbered ports, warm tungsten light.
Agentic Pages Essay 2 May 2026

The two-audience web: why edge interception is how AEO actually gets done.

LLMs.txt is ignored by every major crawler. Robots.txt is a polite suggestion. The only layer that reliably puts machine-readable content in front of AI - and human content in front of humans - is the edge.

11 min read Michael Shaskey
Side-by-side comparison of a Google search ranking first result and an AI chatbot citation list showing different sources
AI Search Essay 1 May 2026

The rank-one delusion: why #1 on Google no longer means cited by AI.

Ahrefs studied 15,000 queries across ChatGPT, Perplexity, and Copilot. Only 12% of cited URLs ranked in Google's top 10. Here is what actually gets you cited instead.

10 min read Michael Shaskey
Terminal window displaying a raw llms.txt Markdown file with sparse white monospaced text on a dark screen, a notebook visible in soft focus behind it.
AI Search Essay 1 May 2026

The llms.txt file: useful fiction, or the next robots.txt?

Three independent studies, 300k+ domains, and 90 days of bot-log data have found no measurable link between llms.txt and AI citations. Here's what the evidence actually says - and the narrower case where the file does earn its keep.

10 min read Michael Shaskey

One letter, every other Friday. Field notes from the agentic web.