Skip to content

Web Presence

The sites linking to you, which of those links AI engines actually use, and how to find outreach targets and clean up bad links.

MonitoringUpdated 26th August 20263 min read

Overview#

Web Presence keeps track of the sites linking to you and how that picture shifts over time.

Links matter for AI visibility for much the same reason citations do. They feed into how an engine works out which sources are worth trusting when it describes your category, and a site that nobody else references gives an engine very little reason to lean on it.

Panels and controls#

Two counts sit at the top: Inbound links (total) and Referring sites. The second is the more honest of the two, since fifty links from one domain amounts to a single relationship while fifty separate domains genuinely is fifty. Gained / lost links shows which way things are moving, and Click a bar to see the links opens up whatever sits behind any given period, which is how you find out what a spike was really made of. Alongside each link, First seen and Last seen tell you whether it is new or has quietly disappeared, and Influence ranks how much the linking site counts for rather than treating everything as equal.

The column worth understanding here is AI-linked, which is particular to this product. It marks the links AI engines are actually drawing on, and that set is always considerably smaller than your total. If you have a large link count with barely anything marked AI-linked, what you have is a broad profile made of the kind of links engines do not use. Anchor and Links to round it out with the text someone used and the page they pointed at.

Inbound links

SourceInfluenceAI-linkedFirst seenAnchor
a review sitehighYes4 months agoproject management for agencies
an industry directorymediumYes2 months agobrand name
a partner blogmediumNo11 months agoread more
a scraper sitelowNo3 weeks agobrand name
Illustrative. AI-linked is the column that separates a link engines draw on from one that only exists.

Links to your competitor(s) picks out sites that link to a rival but not to you, and it is the panel most likely to turn into work you can actually do. The logic is the same as Never cites you over on Citations: somebody already writing about your category, who has already shown they will link out to a company in it, is a far warmer prospect than a site starting from nothing. No outreach targets yet is the empty state, and How to reach out offers guidance on the approach.

Links to your competitor(s)

How to reach out
guidance on approaching each site
No outreach targets yet
the empty state before targets are found
The warmest outreach list you have. Every site here already writes about your category and already links to somebody in it.

Problems worth fixing#

Broken page flags links pointing at pages that no longer exist, and these are pure waste. Somebody made the decision to reference you and the reference lands nowhere; a redirect usually takes a few minutes. No broken links confirms when you have none.

For low-quality links there is Disavowing spammy links and How to disavow, with No high-spam sites as the clean state. Be conservative with this. The screen tells you to Check each domain yourself, which is the right instruction, because disavowal is a blunt instrument and an unfamiliar domain is not the same thing as a harmful one.

Best practices#

  • Referring sites climbing steadily, which is healthier than a spike from one domain.
  • A meaningful share marked AI-linked, since that is the number connecting this screen to everything else in Seerly.
  • Few broken links, given how cheap they are to fix.
  • Overlap with the sources already cited in your category, which is where links and citations start reinforcing each other.

Next steps#

  1. Start with Links to your competitor(s), where the targets are warm and already interested in what you do.
  2. Fix the broken links, the cheapest win on the screen.
  3. Read AI-linked rather than the total. A large number the engines ignore is not the asset it looks like.
  4. Compare gained and lost against your own activity, since a sudden change usually has a cause you can name.
  5. Keep disavowal as a last resort, and check each domain yourself before you use it.

Limitations#

Web Presence describes links, not citations. A site can link to you without any engine ever drawing on it, and an engine can cite a source that never links to you at all. Citations remains the direct measure of the second thing.

Link data also depends on what the underlying crawl has found. A very new link may not appear straight away, and absence from this screen is not proof that a link does not exist.

Related