A 6-Week Plan to Raise Your AI Visibility Score

Getting named by AI tools is weeks of work, not a setting. An answer changes only after the pages the engines read for a question change, and after the engines read those pages again. A plan that runs week by week fits that pace: a few tasks each week, aimed at the pages that matter most, with the score checked against what you expected. This post explains how AEOTrack’s weekly plan is built, what the weeks usually cover, how its estimated score band is worked out, and how to read your real score next to it.

Why weeks, not days

ChatGPT, Gemini, Claude, Perplexity and Grok answer a buyer’s question from the pages they find for it: review sites, directories, “best of” articles, forums and your own site. When you get listed on one of those pages, nothing happens until the page is published, crawled and then picked up again for that question. That takes days at best and often weeks.

The answers also vary between two checks of the same question, so one check moving up or down is not a trend. Every AEOTrack score carries a ± range for exactly that reason (what the ± means). A plan measured over several weeks is the honest way to see whether work is paying off.

How long the plan is

The plan runs for four to six weeks, Monday to Sunday. Its length follows your score when the plan starts: the further you are from being named, the longer it is.

Score at the start of the plan	Plan length
Below 15 (including a first score of 0)	6 weeks
15 to 39	5 weeks
40 and above	4 weeks

When one plan ends, the next starts from your score at that point, so the page always shows the plan that contains this week and says which week you are in.

What goes into each week

The weekly plan is not a separate to-do list. It lays out your action plan, the tasks AEOTrack builds from your own answers, in the action plan’s own order:

1. Get onto the sites the engines read. The pages read most for your questions while the engines named someone else. Each task names the site, the page read most, the questions it was read for and the businesses named from it.
2. Improve the answers you are already in. Pages of yours that already appear in answers, where a clearer answer can move you up.
3. The site checklist. Fixes on your own pages that help the engines read the pages above.

Each week holds up to three open tasks. A task you finish stays in the week you finished it, and its tick is the same tick as on the action plan. Tasks that do not fit in the plan are listed under “later”.

What six weeks usually cover

Your weeks are filled from your own results, so the order follows what your answers show. As an outline, a six-week plan for a site that starts at or near zero usually looks like this. These are also the week titles the free audit shows for a site:

Week	What it is about
1	Fix the issues on your homepage so AI tools can read it easily
2	Get onto the pages AI tools read for your buyers’ questions
3	Answer the buyer questions where other businesses are named
4	Add proof AI tools can quote: reviews, comparisons and answers to common questions
5	Close the gaps on the questions that still do not name you
6	Re-check every question and compare with week 1

For the off-site weeks, how to work each kind of site (directories, review sites, roundup articles, communities) is in the help centre’s off-site plan guide.

The estimated score band, and why it is small

Next to the weeks, the plan draws a shaded band: the score that typically follows if each week’s tasks are done. It is an estimate, labelled as one, and it is not a promise. The engines decide whom they name.

The band starts from your score at the start of the plan and assumes each task adds a small number of points to the 0 to 100 score, counted from the week after it is done, because the engines have to read the changed page first:

• getting onto a page the engines read for your questions: about 0.5 to 2 points;
• improving a page that already appears in answers: about 0.3 to 1.5 points;
• a site checklist fix: 0 to 0.5 points, because it helps the other pages be read rather than moving answers itself.

The numbers are deliberately modest. A first mention moves an average over every answer by a point or two, not by twenty. Past weeks count only the tasks you actually finished; the weeks ahead count the plan as if each week’s tasks get done. A task left undone does not count anywhere.

Reading your real score against the band

The line on the same chart is your real score, the last one measured in each week. Three things help when you read it:

• A single week outside the band is normal. AI answers vary between checks. Look at the line over several weeks.
• Read moves against the ± range. A move inside the range is variation between two checks; a move that clears it is real.
• Citations often come before mentions. Engines can start reading your page as a source before they name you in the text. The per-engine view shows that earlier than the headline number does.

If the line stays flat after the plan, the plan has still told you something: which pages you reached and which you did not. The next plan starts from there.

Who sees what

Account	What the weekly plan shows
Paid plan (Starter, Pro, Agency, Enterprise)	Every week in full: tasks, targets, the estimated band and your real weekly score
7-day free trial	Week one in full; the later weeks as their task counts and bands
Free audit	The week titles for your site, not the tasks

In the app the page is Weekly plan, under Plan in the sidebar, next to the action plan. If your first score is a zero, start with what a score of 0 means and what to do; the plan is the same steps, spread over the weeks.

Mistakes that hold a score down

• Pages that only restate what other sites say. Engines can read dozens of pages on any topic; a page with nothing of its own gives them no reason to name you.
• Treating search rankings as AI visibility. Ranking well in Google does not mean the engines name you; they name the businesses on the pages they read for the question.
• Not measuring. Without a weekly reading you cannot tell whether a change worked, or whether a move was only the answer varying.
• Working on one engine only. Each engine reads and cites differently, so track all five.

Questions

How long is the plan?

Four to six weeks. It is six weeks when your score starts below 15, five when it starts below 40, and four otherwise.

Is the projected score a guarantee?

No. The band is an estimate of what typically happens if the tasks are done. The AI engines decide whom they name, and a single week can sit outside the band.

Can I see the plan on the free trial?

Yes, in part. The 7-day free trial shows week one in full and the later weeks in outline. Every paid plan shows every week.