Week 4 · Days 22 to 30. About 7 hours 30 minutes of focused work, free like the rest of the plan.
Week 4 · Days 22 to 30
Nine days spent where AI decides:on other people's pages.
Every day below carries its own timing, tools, steps, worked example, pitfalls and completion test; the graded evidence behind the plan lives on the full plan page.
Days 22 to 243 hoursEnds with: Ten sent pitches and a response log.
Pitch the pages AI already cites
Listicle outreach is applied week 1: you are not hunting coverage, you are asking to be added to pages the engines already trust.
Have this open
The week 1 shortlistPitch in cited order, most-cited first.
Pitch the week 1 shortlist in cited order, most-cited page first; one inclusion there outweighs five directory listings.
Lead with the number they cannot get elsewhere, dated, and offer it in their house format so saying yes costs minutes.
Send within a day or two of having something fresh; a three-week-old finding is already furniture.
Track pitches like sales: page, owner, date, response, and re-pitch a polite once.
Do it with your AI assistantAI prompt
Role: you are writing outreach asking for inclusion in an article AI assistants keep citing.
Context: the writer runs [article title and URL], cited for [category] questions. My template: [paste the outreach template from the workbook]. My brand: [brand]. Positioning sentence: [paste it]. Two checkable numbers they cannot get elsewhere: [number one, with its date] and [number two, with its date].
Task: personalise the template into a ready-to-send email, plus two subject line options.
Rules: under 140 words; exactly one specific, true sentence about their piece and no other flattery; make declining feel easy; no follow-up threats; the numbers do the selling.
Output: the two subject options first, then the email body.
Worked example · Marnwell Kitchens, an invented brand
Marnwell pitches the regional best kitchen fitters roundup with its 2.4x quote-spread finding and a two-line blurb in the article's own style. The writer adds Marnwell in the next refresh, and that single page begins feeding answers on three tracked questions.
What goes wrong
Pitching a generic editorial inbox instead of the person who wrote the page.
Leading with the company rather than the finding they can use.
Sending one round and never following up, or following up four times.
You are done with days 22 to 24 when
Ten pitches are sent, in cited order.
Each carries a dated number the writer cannot get elsewhere.
Every pitch is logged with its owner and date for follow-up.
Days 25 to 261 hourEnds with: A running, triggered review request sequence.
Push review volume on G2 and Capterra
Review velocity is a habit dressed as a campaign: the ask must live at the moment of success, not in a quarterly blast.
Have this open
Your review platform profilesComplete them with the canonical sentence from day 5.
A trigger moment in your deliveryThe point where the customer feels the win.
How to do it
Anchor the ask to the trigger moment your product or service actually has, and send it from a named person.
Ask for honest and specific, never positive: reviews that name a use case are the ones answers quote.
Aim at steady count, not score defence; the measured correlation runs through volume, and a 4.6 with 80 reviews beats a 5.0 with 6.
Do it with your AI assistantAI prompt
Role: you are a customer marketing lead writing a review-request sequence.
Context: brand [brand]; platform [G2 or Capterra]; sender [named person]; trigger moment [a success moment in your product, e.g. their first report ships].
Task: write three messages: 1) a 60-word ask sent at the trigger moment; 2) one gentle follow-up a week later; 3) a short thank-you once a review appears.
Rules: ask for an honest review, never a positive one; no incentives beyond what the platform allows; suggest the reviewer mention their specific use case, because specific reviews are what AI answers quote; no guilt in the follow-up.
Output: the three messages, labelled 1 to 3, nothing else.
Worked example · Marnwell Kitchens, an invented brand
Marnwell's ask goes out the day each kitchen is signed off, from the fitter who did the work. Reviews mentioning small galley kitchen and appliance fitting start surfacing in answers to exactly those niche questions.
What goes wrong
Asking for a positive review rather than an honest one, which the platforms and buyers both punish.
Running a quarterly blast instead of an ask tied to the moment of success.
Offering an incentive the platform forbids, risking the whole profile.
You are done with days 25 to 26 when
Profiles are complete and carry the canonical positioning sentence.
A three-message sequence is live at the trigger moment, from a named sender.
The first asks have gone out.
Days 27 to 282 hoursEnds with: Real thread participation and two pitched angles.
Show up where the threads are, by hand
Community presence and PR both trade on the same asset: a named person with something checkable to say.
Have this open
The cited threads from week 1Only the threads the engines actually quote.
A named account with historyA fresh account praising one brand is spam.
One publishable findingA single quotable sentence with the brand inside it.
How to do it
Enter only the threads where you can answer the actual question first and mention any product second, or never.
Use one named account with a real history; a fresh account that only praises one brand is spam whatever it types.
Pitch the finding, not the firm: one plain quotable sentence with the brand named inside it is the whole ask.
Do it with your AI assistantAI prompt
Role: you are a community strategist who despises spam.
Context: AI assistants cite these threads and questions in the [category] market: [paste them]. The person who would participate: [named person] from [brand]. A finding we can publish: [paste one number or result].
Task: 1) for each thread, judge whether [named person] can add real value as a named participant, yes or no with one reason; 2) for each yes, draft a two-sentence opening that answers the actual question before any product is mentioned; 3) turn the finding into two digital PR angles.
Rules: each PR angle is a single plain sentence a journalist could quote with [brand] named in it; never draft anything that opens with a product pitch.
Output: the thread judgements, the openings, then the two angles.
Worked example · Marnwell Kitchens, an invented brand
In a thread the engines cite on kitchen fitting costs, Marnwell's fitter answers with the measured quote spread and no link at all. A trade publication picks the figure up a week later, crediting Marnwell Kitchens in text, an unlinked mention that still counts.
What goes wrong
Posting from a new account created for the purpose, which reads as astroturf.
Answering with a pitch instead of an answer, which gets removed and remembered.
Pitching the company story when only the number is publishable.
You are done with days 27 to 28 when
At least two cited threads carry a genuinely useful answer from a named person.
Two PR angles exist as single quotable sentences naming the brand.
Both angles are pitched to real people.
Days 29 to 301 hour 30 minutesEnds with: The day 30 report and a booked operating rhythm.
Re-run the baseline and set the rhythm
The re-measure is the plan grading itself, and the honesty of the report is what makes the next 30 days plannable.
Have this open
The day 1 baseline and screenshotsThe same instrument, twice.
Re-run week 1 exactly: same questions, same platforms, same sheet, and only the no-brand questions in the headline numbers.
Attribute movement to causes only where a date lines up with an action; everything else is weather.
Set the standing rhythm before closing the report: one fan-out review, one refresh, one outreach ask per week, owned by a name.
Do it with your AI assistantAI prompt
Role: you are an analyst writing the honest day 30 report.
Context: day 1 and day 30 AI visibility baselines for [brand] are pasted below. Stable citations typically take 8 to 16 weeks or longer, so day 30 measures direction, not victory.
Task: report 1) share of voice change per platform, counting only the questions that never contained our name; 2) which question groups moved; 3) which competitors gained; 4) what did not move at all; 5) the three highest-leverage actions for the next 30 days.
Rules: flat numbers get reported flat, no spin; tie every claim to a number in the data; where the data cannot support a conclusion, say so instead of concluding.
Output: five numbered sections, under 500 words.
Data: [paste both tracking sheets].
Worked example · Marnwell Kitchens, an invented brand
Marnwell's day 30: ChatGPT mentions up from 4 to 9 of 50, Perplexity still at 1, the corrected appliance-fitting error gone from every platform. The report says plainly that the listicle inclusion did more than all on-site work combined, and books next month accordingly.
What goes wrong
Letting brand-name questions into the headline number, which flatters and misleads.
Claiming credit for movement no dated action explains.
Closing the report without naming who owns the weekly rhythm.
You are done with days 29 to 30 when
The week 1 questions have been re-run on the same platforms.
Deltas are reported on no-brand questions only, flat numbers included.
Three next actions and a weekly rhythm have owners and dates.
What this week leaves youYou end day 30 with a measured delta, a working rhythm, and the honest expectation set below.
The workbook
The workbook turnsreading into doing.
A printable day-by-day checklist, the 48-prompt starter pack organised by buyer intent, the week 4 outreach template and the baseline tracking sheet. Email one carries the workbook link; seven more follow, paced to the plan’s weeks across thirty days, then the newsletter’s occasional notes.
Questions
Asked about this week, answered straight.
Including the ones where the honest answer costs a sale.
What if listicle writers ignore my pitch?
Most will, and the maths still works: the shortlist has ten pages and one inclusion pays for the week. Improve odds by pitching people not inboxes, leading with the unpublishable-elsewhere number, and making the addition easy in their format. After one polite follow-up, move down the list; a page that ignores you in August often refreshes in October.
Are paid listicle placements worth buying?
Rarely for AI visibility. Pages that sell placement tend to list everyone, and directories that list everyone are cited as directories, not as recommendations. The measured signal runs through editorially chosen mentions. Where a paid placement also carries genuine editorial standards and real traffic, judge it as advertising on its own merits, not as citation strategy.
What counts as a good day-30 result?
Direction on the no-brand questions, corrections landed, and the first third-party inclusion in flight. Typical honest outcomes look like single-digit share of voice becoming low double digits on one or two platforms while others sit flat, because stable citation patterns take 8 to 16 weeks or longer. A flat day 30 with the foundations fixed and outreach live is on schedule, not failing.
Every step of this week, with the graded evidence behind it, sits in the full 30-day plan.
Start here
See where you show up in AI answers today.
Add your site and RankX AI suggests the prompts to track, monitors the keywords that matter and audits your pages, with your first results minutes after you finish setup.