Other tools stop at the dashboard.
Referenced hands you the fix.
After every scan, Referenced turns what it found into one ranked list of work: the page to write, the thread to answer, the publication to pitch, how AI describes you, and anything blocking AI crawlers.
7-day trial, no card
Write "choosing your first ultra shoe"
You are absent from nine answers that name two rivals. One page covers all nine.
Allow answer crawlers in robots.txt
Your robots.txt blocks a crawler that answer engines use, so they cannot read your pages.
Answer the returns thread
Cited on five negative answers and still unanswered. Draft a reply to edit.
Answer the returns criticism
Engines keep saying returns are slow. Say how your returns work, plainly, on your returns page.
Pitch runnersworld.com
Cited on 21 answers, never beside you.
Answer the sizing theme
Five answers say the shoe runs small. Say plainly how it fits.
Four kinds of action
Your own site, third-party pages, how AI describes you, and technical access.
Replies and pitches drafted
Ask for a reply to a thread or a pitch to a publication, written in your context and ready to edit.
Nothing auto-posted
No account connections to post on your behalf. You read it, you change it, you send it.
Refreshed after every scan
The list is ranked again after every scan. An action the AI stops suggesting closes after two scans in a row, and one whose gap is gone closes straight away.
What is GEO?
Generative engine optimisation is the work of getting named in the answer: being on the sources engines quote, answering plainly on a page they can use, and being described in terms a buyer wants to hear.
Not SEO with a new label
There is no ranking to climb and no page two to be relegated to. There is one answer, it names two or three brands, and you are in it or you are not.
What a recommendation is made of
Numbers first, sentence second. Every recommendation shows the facts it was computed from, so you can disagree with the advice without doubting the arithmetic.
Write "choosing your first ultra shoe"
Hoka and Merrell are named in all nine answers to this question across three engines. You are named in none of them, and no page of yours is cited on any of them.
One list, ranked by answers affected.
Every action shows how many of your stored answers it affects, so the biggest job is at the top whatever kind it is. Each plan keeps a set number of actions open per site: 5 on Brand, 10 on Pro and 15 on Agency.
Your own site
Pages to write or improve where rivals are named and none of your pages is cited.
Rivals are named in answers that do not cite you
Third-party pages
Threads to answer and publications to pitch that engines already cite.
A cited page never names you
How AI describes you
Unflattering claims that keep recurring, with an honest response to each.
The same claim repeats across enough answers
Technical access
A robots.txt rule that blocks a crawler answer engines use.
An answer crawler is blocked from your site
Four kinds of work, ranked by what it is worth.
Each kind comes out of the same scans and is ranked on the same scale, by how many answers it affects. Technical access, such as a robots.txt rule that blocks an answer crawler, joins the same list.
Where dashboards stop.
Most tools in this category measure well and end there. The difference is not more charts; it is that the chart hands over a job with the reasoning attached.
Recommendation questions, answered.
Where they come from, how often they change, and what stays your decision.
Does it post or publish anything for me? + −
No. Every draft (a reply or a pitch) is yours to read, edit and send. Nothing is posted automatically, and nothing is written to your site.
How often are recommendations refreshed? + −
After every scan. The list is ranked again. An action the AI stops suggesting closes after two scans in a row, and one whose gap is gone closes straight away, so the list keeps up with what is true.
Where do the recommendations come from? + −
They are built from the answers stored in your scans: who is named, which pages are cited, how AI describes you, and whether your robots.txt blocks an answer crawler. Each action shows the facts it was computed from. The facts are computed first and only the wording is written.
What counts as a content gap? + −
A prompt where a competitor is named in every answer and you are named in none of them, weighted by how many rivals and how many answers are involved, not a single missed mention.
How many actions do I get? + −
Each plan keeps a set number of open actions per site, the highest ranked first: 5 on Brand, 10 on Pro and 15 on Agency. Each one shows the numbers it was computed from.
Stop reading dashboards. Start closing the gaps.
Free 7-day trial, no card: your first action list lands with your first scan.
Feature
Visibility
Your score across every AI engine.
Feature
Sentiment
What AI thinks of your brand.
Feature
Competitors
Share of voice, brand by brand.
Feature
Citations
Which pages AI actually cites.
Feature
AI Traffic
GA4 + Search Console, overlaid.
Feature
Recommendations
The fix for every gap, ranked.