October 17, 2026

A Complete Guide to the AEOTrack Action Plan Page: The Sites AI Reads First

Introduction

The Action plan (Improve → Action plan in the sidebar) tells you what to do next, in the order the evidence says: the sites AI reads first, then the answers you're already in, housekeeping last.

It is built from your real answers: which pages each engine read when it answered your questions, and who it named. This guide walks through the page using a fictional example company, Acme Coffee Roasters (acmecoffee.example). The how-to for each kind of site, with the evidence behind the approach, is in Get named: work the off-site plan. To see the same work laid out week by week, with an estimated score band and your real weekly score against it, see Your 6-week plan.

The three sections

Section	What is in it	Why it comes here
1. Get onto the sites AI reads	Up to 10 sites the engines read in the last 30 days when they answered your questions without naming you, and never when they did	AI engines name the brands mentioned on the sites they read. This is where most new visibility comes from.
2. Move up in the answers you're already in	Tasks drawn from answers where the engines already read your pages: statistics, quotes and sources for your pages AI reads, short pages that are cited	Improves answers you are already part of.
3. Hygiene checklist (collapsed)	Schema, llms.txt, meta tags, titles, headings, canonical tags, alt text, internal links, robots.txt	Worth doing once; tests found no measurable effect of schema or llms.txt on whether AI names you.

1. Get onto the sites AI reads

Which sites, in which order

A site is listed when the engines read it in the last 30 days while answering your questions without naming you, and never read it in an answer that did name you. Your own site, your tracked competitors' own sites, search engines and .gov/.edu sites are left out. YouTube, Wikipedia and LinkedIn count; X, Facebook and Instagram don't, because the engines rarely cite them.

The sites are ordered by how often each was read, weighted by engine. A read in an answer that named a competitor counts in full; a read where no competitor was named counts half. A site read only once with no competitor named is left out. The top 10 are shown, each with its rank and the kind of site (review site, list article, Reddit, YouTube, Wikipedia, press, LinkedIn or other).

What a task shows

An off-site task in the Action plan for the fictional Acme Coffee Roasters: get onto roasterratings.example, read 71 times in 30 days for "best light roast coffee brands", with the engines that read it, the rivals named instead, the playbook steps and the status buttons. Task 1 for Acme Coffee Roasters (fictional demo data): roasterratings.example was read 71 times in 30 days for "best light roast coffee brands" while the engines named Bluebird Coffee and Northwind Roasters.

The question, with the engines that read the site for it, and the other questions it was read for.
Which engines read it, with a count for each (in the example: Gemini 16, Claude 15, Grok 15, Perplexity 13, ChatGPT 12).
Named instead: the competitor the engines named.
The cited page, linked.
Read N times in 30 days.
How to do it: the steps for that kind of site. They tell you what to do, never promise a result, and include the honesty rules (say who you work for, follow each site's rules, no paid or self-written reviews).

Statuses

Each task carries To do, Contacted, Listed or Not possible, and the section can be filtered by status.

Listed is not reopened for 30 days. An open task also closes itself as Listed once an answer that names you reads that site.
Not possible (a forum that bans vendors, another company's own page) stays closed.

Download outreach list (CSV)

On Pro and up, Download outreach list (CSV) exports the sites with the columns site, url, source_type, engines, questions, rivals_named, times_read, suggested_action and status. It contains no personal contact details; finding the right person at each site is your step. On Starter the button shows a lock and links to Billing.

2. Move up in the answers you're already in

Add statistics, quotes and sources to the pages AI already reads. The task lists your own pages the engines read, and for which questions. These suggestions only appear for pages AI already reads.
Short pages are flagged only when the engines cite them.
Other tasks drawn from your answers, for example questions where the engines read your page but name someone else.

Progress and backlog

Below the first two sections:

Summary cards: Website Health (the site-structure checklist score, e.g. 78%), Open improvements, Completed, Quick Fixes (open hygiene fixes estimated at 10 minutes or less; quick to do, with no promise about what they change) and Last scan.
"Is the plan working" plots the tasks you completed against your AI visibility score, so you can see whether the work lines up with a rise. Example (fictional): Acme Coffee Roasters closed three tasks on 14 Aug, 28 Aug and 11 Sep 2026; its score was 11, then 15, then 27 at each one.
Backlog: open, in-progress and completed counts, and where the open work comes from.
Where to spend your effort: your questions by whether AI names you and whether it cites your site (strongholds, visible not cited, cited not visible, needs attention).

Screenshot of the AEOTrack Action Plan summary cards, the "Is the plan working" chart and the backlog. The summary cards, the "Is the plan working" chart and the backlog.

3. Hygiene checklist

Collapsed at the bottom of the page, with one line of explanation: basic housekeeping for your pages, worth doing once, but tests found no measurable effect of schema or llms.txt on whether AI names you. Open it to see each item and its steps.

Push to GitHub opens a pull request with a paste-ready snippet, or an llms.txt placed where your framework serves it. Nothing existing is replaced, and the FAQ item can only be pushed once you have approved FAQs. See the GitHub integration.
Mark as done closes an item you fixed yourself. Completing an item also submits the pages it touched to IndexNow, if you have set it up.

Step-by-Step Workflow

Open section 1 and start with the top site.
Open the cited page and read what it says about your competitors.
Follow "How to do it" for that kind of site, then set the task to Contacted.
Set Listed when you are on the page, or let it close itself when an answer that names you reads that site. Mark Not possible when there is no way on.
Repeat for three to five sites a week, then check the next weekly results email for questions that moved.
Do the hygiene checklist once, when you have a spare hour.

Frequently Asked Questions

How often does the plan refresh? It is rebuilt weekly after your scheduled checks, from the last 30 days of answers.

How is the order decided? Section 1 by how often each site was read, weighted by engine (reads where a competitor was named count in full, others half); section 2 by how many answers each task rests on; the hygiene checklist in a fixed order. Where a task has a count, its card shows it, for example "Read 71 times in 30 days"; a task without a count shows no label. A task shows no points and no forecast of what it will change; the estimate in Your 6-week plan is for the six weeks of tasks as a whole, and it is an estimate, not a guarantee.

Why isn't my competitor's own website on the list? Tracked competitors' own sites are left out of section 1: there is no listing to get on them.

Why is a site I got onto still listed? Mark it Listed. It won't reopen for 30 days, which gives the engines time to read the page again.

Conclusion

Work the plan from the top: get onto the sites the engines read, strengthen the pages they already read, and keep the hygiene checklist for a quiet afternoon.