Tag
Start here
Classify thousands of queries by intent with a local model, including the cascade design that works around Laya's twenty-option limit, the confidence threshold you set from your own data, and how to measure whether it is actually right.
Tag
Most teams analyze a sample of Search Console data because a full export is too much to sort by hand. Here is a workflow that classifies the whole export with Jev, plus the three traps that will quietly produce confident wrong answers.
A century-old water brand went from missing on brand decision prompts to the top three AI recommendations. Inside the intent map, semantic rebuild and source plan behind the reported 12% to 38% shift.
AI Overviews do not hit every query the same way. Informational queries lose 70-80% of clicks, commercial queries 35-45%, transactional 20-30% — and branded queries actually gain. A query-type survival matrix with a concrete move for each of the four intents, so you defend the right ones and let the others go.
Learners now ask AI assistants which prep course is good before they open a browser. How exam platforms get recommended and cited: the answer anatomy, the quotable assets, the freshness rules and a 30-day plan.
The best time to run a landing page audit is before launch, before paid traffic, and before a redesign. Check the three info gaps — audience, value, action path — before you rewrite anything.
Get the codex-seo-sxo skill, read a Google SERP backwards for the page type it rewards, and watch it give one page two verdicts in one day: ALIGNED on its commercial keyword, MISMATCH on its informational one.
Point Codex at any URL and get a five-category page diagnosis: on-page elements, content signals, technical meta, schema, and images, with a score card and a prioritized fix list.
Why AI Overviews Skip Tool Keywords — The 2026 SEO Playbook Backed by Data When an AI Overview shows up above your result, Google pushes the first organic listing down by roughly 1,674 pixels — past t...
Keyword clustering groups keywords with the same search intent so you can map each cluster to one page. Here are the 2026 methods, free tools, and a copy-paste Codex skill that does the work for you.
Commission a bounded Workbubby research deliverable that turns source files, customer language, and SERP notes into a keyword recommendation your team can accept or reject.
Use Claude Code to connect keyword evidence to the pages, templates, and constraints already present in a website repository, then prepare a reviewable SEO change request.
Learn how to use Hermes Agent, the open-source AI agent framework from NousResearch, to run a supervised weekly keyword research workflow. Set up exception rules and evidence checks once, then review a small decision-ready queue every week instead of starting keyword research from scratch.
