Linkology (formerly UK Linkology)

22/10/2026

22/10/2026

Could we finally get a clearer picture of how much website traffic comes from Google's AI answers?

That's the question behind a recent report about tracking parameters appearing on links inside AI Overviews and AI Mode.

The idea is interesting: if AI-generated answers carried a distinct referral tag, businesses could separate those visits from traditional organic search traffic.

But the original observation has an important complication.

The `st_source=ai_overview` parameter has not been confirmed as an official Google experiment, and subsequent reporting traced it to a browser extension.

The underlying measurement problem hasn't disappeared.

Google's Search Console provides AI-specific impression reporting, but it still doesn't offer a dedicated click breakdown for those AI experiences.

For businesses investing in content and AI visibility, understanding which efforts actually generate qualified visitors and leads remains essential.

If you'd like help understanding your brand's AI visibility and where opportunities may be missing, book a free call with us.

21/10/2026

6 days.

Page one and the AI answer are related, but they aren't the same thing. The brands that get recommended earn a place in both.

20/10/2026

Can AI find your website? Great.

But can it understand who you are, why you're credible, and what you actually know?

Those are very different questions.

An AI readiness audit of 50 major websites, discussed in Search Engine Journal, reported just 38% compliance for attribution and meaning.

The research highlights three important areas:

🔎 Retrievability: Can AI systems access your content?

Attribution: Can they identify the source and its credentials?

Meaning: Does your content communicate genuine knowledge rather than generic information?

For businesses investing in AI search visibility, this is a useful reminder that technical improvements are only one part of the picture.

Try reviewing three of your highest-traffic pages this week.

Check whether authorship is explicit, important claims are supported, and the content offers something more substantial than a basic definition.

The objective is to make your expertise easier to identify, understand, and verify.

20/10/2026

7 days.

Being found now means being in the answer, not just on the page. The brands that get recommended are the ones the rest of the web already talks about.

19/10/2026

8 days.

AI looks for agreement. When trusted sites tell the same story about your business, you become easier to recommend.

19/10/2026

Your older listicles might need more than a new publication date.

Kevin Indig's research into listicle performance found that freshness was one of the strongest consistent signals associated with top-three rankings.

The research covered nearly 60,000 search queries, with a closer examination of 3,000 ranking listicles.

And here's the useful takeaway:

Simply changing a timestamp isn't the same as keeping an article relevant.

If your listicle recommends tools, services, products, or providers, ask yourself whether those recommendations still hold up today.

Start with your 10 highest-traffic listicles.

Check whether search engines can interpret their dates, review what has changed over the last 6–12 months, and replace entries that no longer deserve their place.

Then document the changes.

💡 The objective isn't to convince Google that your content is new. It's to make sure readers are still getting recommendations they can trust.

18/10/2026

9 days.

One mention in a publication your buyers already trust can travel further than a year of blog posts. Earned coverage is what gets remembered, by people and by AI.

17/10/2026

10 days.

A search page gives ten options. An AI answer usually names a handful. Fewer spots means every recommendation counts for more.

16/10/2026

11 days.

Buyers trust what others say about you more than what you say about yourself. AI assistants work the same way: they recommend the brands that trusted sources already talk about.

15/10/2026

12 days to go.

Your next client has probably already asked an AI assistant who to hire, before they ever searched for your name. Over the next 12 days, one shift a day in how buyers choose who to trust. On the 27th, what we've built for it.

23/09/2026

OpenAI is testing a very different kind of ChatGPT ad.

Instead of clicking an ad and immediately being sent to a website, users can choose to start a conversation with a business-sponsored AI agent.

💬 They're called Sponsored Agents.

Users can ask follow-up questions, explain what they're looking for and explore a business's products or services before taking the next step.

OpenAI says these agents are clearly labeled and their conversations are separate from ChatGPT's independent answers.

For businesses, the interesting part is what happens to the traditional advertising funnel.

An ad click doesn't necessarily have to become a website visit anymore.

It could become a conversation first.

That means brands may need to think carefully about the product, service and contextual information these agents use, as well as how they measure whether those conversations actually generate qualified engagement.

18/09/2026

This is Mission: SPACE at EPCOT.

They strap you in, launch you towards Mars and, thankfully, bring you safely back to Florida. 🚀

But it got me thinking about one of the questions businesses keep asking us:

How can we use the growth of AI search to generate more visibility?

There's a big opportunity here.

People are increasingly using AI-powered search to research options and discover businesses, while platforms like ChatGPT can recommend brands directly within their answers.

But simply getting visibility isn't enough.

A rocket without a guidance system is just an expensive firework.

You need the boost and the direction.

That means building the authority, trust, and visibility signals that help AI systems understand your business and give them reasons to recommend it.

We've already seen one client nearly triple their share of voice in a single month.

That's the kind of liftoff we're working towards.

Book a free call and let's see what's possible for your business.

18/09/2026

Google has introduced another way for publishers to gain prominent visibility inside AI Search.

Developing-topic article carousels are now live in AI Mode after Google introduced the format in AI Overviews.

For some searches around developing topics, users can see a prominent row of timely articles directly within the AI-generated response.

Google says the goal is to connect people with original coverage and a range of perspectives.

The carousels can also highlight publications that individual users have selected as Preferred Sources.

For brands and publishers, this creates another interesting inclusion opportunity.

It's no longer simply about whether your content ranks traditionally or earns a citation inside generated text.

Google is creating increasingly prominent surfaces for original content within the AI experience itself.

That makes building topical authority, producing genuinely useful original coverage, and developing an audience that actively seeks out your content even more important.

17/09/2026

Google has finished rolling out its Generative AI performance report worldwide.

And for SEO teams trying to measure AI visibility, this is an important step forward.

Search Console can now show impressions from AI Overviews and AI Mode, with visibility broken down by:

📄 Page
🌎 Country
📱 Device
📅 Date

There is still an important limitation.

The report doesn't provide AI-specific click data.

So this isn't a complete attribution solution, and an impression shouldn't automatically be treated as traffic or engagement.

But it does give businesses a native Google baseline for understanding which pages are appearing across its generative AI search experiences and how that visibility changes over time.

If you'd like to understand how we attribute AI visibility beyond the data available in Search Console, get in touch.

17/09/2026

I'm at EPCOT's World Showcase.

11 countries. One lagoon. And, strangely enough, a pretty good representation of what search looks like today.

Businesses still have one goal:

Be visible.

But there are now many different places where that visibility needs to happen.

ChatGPT, Gemini, Perplexity, Google AI Overviews and traditional search each operate differently.

🗺️ Think of them as separate territories.

Being established in one doesn't automatically give you visibility in another.

You could have a strong position in traditional Google Search while barely appearing in AI-generated answers elsewhere.

That's why we're moving beyond thinking about visibility as simply “Where do we rank?”

The better question is:

“Where are our customers discovering businesses, and are we present there?”

We've spent 15 years helping businesses navigate search.

AI is another territory, and we've got the map.

Book a free call and let's see where your business needs greater visibility.

17/09/2026

Google has started automatically expanding some AI Overviews into much fuller answers.

Previously, a user might see a shorter AI Overview and need to click “Show more”.

Now, for certain queries, Google can expand the response automatically and display the “Ask anything” prompt underneath it.

That means traditional organic results can be pushed significantly farther down the page.

Google says the expansion happens when its systems determine that a fuller response will be useful, although it hasn't disclosed how frequently this experience appears.

For SEO teams, the implication is important:

Ranking position and actual search visibility are becoming increasingly different things.

If more of the initial search experience is occupied by AI-generated answers, monitoring whether your brand is mentioned or cited inside those surfaces becomes increasingly important too.

17/09/2026

Most white-label link deliveries look perfectly fine on a spreadsheet.

Then you open the page.

Read the anchor context.

Check the surrounding copy.

And suddenly, the placement doesn't look much like an endorsement.

It looks like filler. 👀

Google confirmed the August 2026 spam update on 18 August, rolling out globally across all languages, according to Search Engine Land.

For agencies working with delivery partners, that's another reason to look beyond whether a link simply exists.

Before you invoice, we'd run three checks:

1. Does the paragraph around the link make sense without it?

2. Do the surrounding brand and topic mentions actually match the client's category?

3. Would you still approve the page if the link was removed? 🔍

If the answer to any of those is "not really," we'd question what the client is actually paying for.

Because a placement shouldn't need the spreadsheet to make it look valuable.

The page itself should make the case.

What's the first thing you check when a supplier sends over live URLs?

17/09/2026

You don't necessarily need a huge website to earn citations in AI answers.

But you do need to make it clear what your page actually answers.

One study looked at 1,200 AI answers and found smaller sites being cited regularly too.

That led us to run a quick audit on some of our own pages.

The issue became obvious.

Many pages were originally written for humans skimming a screen, so their H1 might say something broad like:

“Our SEO Services”

But an AI system looking for a source is often trying to answer something much more specific.

🔎 So here's a quick audit you can run:

Pick one of your important pages.

Read the H1.

Identify the specific question or buying situation that page should answer.

Then check whether those two things clearly align.

If they don't, you've potentially found a useful improvement that can take just 5–10 minutes to make.

16/09/2026

I'm standing inside Club Cool at EPCOT, where Coca-Cola gives away drinks from around the world for free.

Which raises an obvious question:

Why would a company worth hundreds of billions give its product away?

🥤 Because experiencing something firsthand can be far more convincing than being told how good it is.

You experience the value before anyone asks you to buy.

That's exactly how we think businesses should experience our service too.

Every business that comes to us can start with a free AI Visibility Blueprint.

We'll show you where you're currently visible in AI search, where the gaps are, and what we'd focus on fixing before you spend anything with us.

You get to see how we think and experience the client journey from day one.

We've got 15 years in the sector, backed by a money-back guarantee.

And the first sip is still free.

Book a free call and we'll get your AI Visibility Blueprint flowing.

16/09/2026

Your business has three names online.

Google sees a fourth.

AI models pick a fifth. 😬

This is how a real founder might describe a local visibility problem that looks minor until it starts affecting how the business is presented.

Your registered company name says one thing.

Your website header says another.

Your Google Business Profile is slightly different again.

Then your invoices use another variation.

Search Engine Roundtable reported on 12 August 2026 that Google is showing AI-generated summaries above some local results.

Suddenly, that inconsistency isn't just a tidy admin issue.

It's potentially shaping how your business gets described in the answer sitting above the map pack. 📍

The fix is refreshingly simple:

✅ One legal name
✅ One trading name
✅ One address format

Then use them consistently everywhere.

Your website. Google Business Profile. Directories. Invoices. Social profiles.

Give AI a clean source of truth to work from instead of making it choose between five different versions of your business.

Where does your business name change between your website, Google Business Profile and invoices?

16/09/2026

Before increasing spend on ChatGPT ads, there's an audit worth running.

Don't just look at impression volume.

Look at relevance.

Industry data suggests some ads are appearing alongside commercial prompts that have little topical connection to what they're advertising.

That creates an important distinction:

📈 Impression volume tells you how often you're appearing.

🎯 Targeting relevance tells you whether those appearances make sense.

Pull your active campaigns and test a sample of commercial prompts relevant to your category.

Record what appears alongside each prompt and flag any placement where the query and the ad topic don't match.

Even a relatively small sample can give you a better idea of whether your spend is landing in conversations that actually make sense for your business.

16/09/2026

This is an interesting one.

Google is testing a program that could pay publishers when their content contributes to its AI-generated answers.

The AI Contribution pilot covers Gemini, AI Overviews, and AI Mode.

💰 Publishers taking part can see monthly earnings inside Search Console, although exactly how Google calculates those payments remains unclear.

Why does this matter?

Until now, much of the conversation around being cited in AI answers has focused on visibility, traffic, and brand recognition.

If this program expands, there's potentially another piece to consider:

Direct revenue.

It's still an early-stage pilot, so there's plenty we don't know yet. But this could become an important development for publishers and anyone creating content for AI search.

🔗 Read the source:
https://www.searchenginejournal.com/google-tests-paying-publishers-for-ai-answers-via-search-console/589414/

15/09/2026

Standing underneath EPCOT's famous Spaceship Earth, it's pretty easy to feel small.

It's 180 feet tall.

And that feeling isn't too different from what many business owners are experiencing with search right now.

AI is changing how people discover businesses. Traditional website traffic is under pressure. More of the customer journey can happen before someone ever clicks through to your site.

It can feel like the landscape changed overnight and your business got smaller.

But your size isn't necessarily the problem.

Your visibility is.

🔎 You don't need to be the biggest company in your market.

You need to make sure you're being found in the places where people are now searching, researching, and getting recommendations.

We've spent 15 years helping brands become harder to ignore through search. Now we're doing the same as discovery moves into AI.

If you'd like to see where your business stands, book a free call.

Let's get you found.

15/09/2026

Your client sees their competitor named in a ChatGPT answer.

They ask why it wasn't them.

What do you show them on Monday? 👀

Most white-label reports still lead with DR gains and referring domain counts.

That was fine two years ago.

But it doesn't answer the question the client is actually asking now:

"Why is AI recommending them instead of us?"

An iPullRank analysis reported by Search Engine Land this week placed ChatGPT as the sixth most-clicked destination from Google Search, ahead of most traditional publishers.

Your client's buyers aren't staying on one platform anymore.

They're moving between Google, ChatGPT and branded search in a single session.

📊 If your retainer report can't show where the client's brand appears across that journey, someone else's eventually will.

So start logging AI mentions the same week you deliver links:

🔍 Prompt sets
📅 Dates
🤖 Models
🏷️ Which brands were named

Then track how that changes over time.

Links still matter.

But clients are increasingly going to ask whether all that work is helping them become the brand AI actually names.

That's the layer they'll notice.

What are you adding to your Q4 reports?

15/09/2026

Google ranking volatility spiked again around August 12–13.

That doesn't automatically mean there was a confirmed algorithm update.

And it definitely doesn't mean your first move should be reverting content or changing links.

📊 Before you act, work through the evidence.

First, check the timing. Did the decline begin before, during, or after the volatility spike?

Second, compare query groups. A sitewide decline tells a very different story from a drop affecting only a few topics.

Third, audit your own activity. Were there technical changes, deployments, or content updates during the same period?

Rule those out before blaming Google.

The best next step is simple: pull the ranking data, note what moved and what held, and make decisions from that rather than reacting to the noise.

14/09/2026

Every reseller sells "quality links."

Very few can actually prove it. 🔍

And that's where agency partnerships either compound over time or quietly start to fall apart.

Search Engine Journal reported this month that Google research found frontier AI models fail to recall 26 to 34% of facts they encode, with recall dropping further when the subject-object order flips.

For agency partners, there's an important takeaway:

Consistency of evidence around a client's brand now matters just as much as the links themselves.

If your delivery partner can't show you:

📋 The publisher vetting logic
👀 The manual review process
🧩 How client entity signals are protected across placements
..then you may simply be buying volume dressed up as quality.

And clients tend to feel that difference pretty quickly.

Here's a simple test.

Ask your partner to show you their last ten placements and explain the reasoning behind each one.

Why this publisher?

Why this angle?

Why this placement for this particular client?

Their answers will tell you far more about "quality" than a sales page ever will.

What does your quality control actually look like on paper?

11/09/2026

Here's an interesting relationship from the analysis:

A 0.334 correlation between branded search demand and AI citation frequency.

It's important not to overstate what that means. Correlation isn't causation, and 0.334 isn't an enormous relationship.

But there is an interesting detail.

📊 It was meaningfully stronger than several classic technical signals.

That gives us another reason to think beyond the website when talking about AI visibility.

Technical SEO matters.

But brand recognition, branded demand, and the wider signals surrounding a business may be part of the picture too.

🎥 See how this connects to entity clarity:
https://youtu.be/DESQwJFkoCc

11/09/2026

Scaling topical authority across ten client sites is a very different problem from building it on one.

Most agencies try to solve it by adding more writers.

But more people rarely fix a system problem. ⚙️

The real issue is your entity register.

If each client's core entities, aliases and relationships live in one controlled source, briefs get faster, writers make fewer assumptions, and gaps in coverage become much easier to spot.

Microsoft Clarity's new Topic Insights now groups AI citations by subject and shows where a brand's topical coverage is thin, according to Search Engine Journal on 10 August.

That's more than an interesting report.

It's a diagnostic backlog waiting to become cluster work. 🔍

Build the register once for each client.

Canonical name. Aliases. Category. Related entities. Supporting URLs.

Then reuse it in every brief.

Your writers stop guessing, your editors have something consistent to check against, and delivery quality is far less likely to drift as more accounts come on board.

📈 Scaling isn't just about increasing output.

It's about building a system that still works when you're managing fifteen clients instead of three.

What tends to break first when you scale?

10/09/2026

A PR mention comes in, but there's no backlink.

Does that mean it has no value?

That's based on an older model of how search works.

AI systems can build their understanding of a business partly from how independent sources describe it across the wider web.

That means the context surrounding your brand matters too.

What category are you associated with?

Which services appear next to your name?

What expertise are you repeatedly known for?

🔗 A backlink still helps. But it's no longer the only part of the mention worth evaluating.

Your website tells machines who you say you are.

Third-party coverage helps reinforce how the wider web understands you.

🎥 See why this matters for AI search:
https://youtu.be/DESQwJFkoCc

10/09/2026

Most white-label link reports get rejected for one surprisingly simple reason.

The link looks fine.

The paragraph around it doesn't. ⚠️

SEOFOMO's 9 August roundup highlighted Daniel's review of Google patents and testing around page segmentation and link placement.

The practitioner conversation has moved on.

It's no longer just about how many links you place. It's about whether the surrounding paragraph actually supports the claim being linked.

That changes what quality control needs to look like.

If your delivery team is checking DR and anchor, then calling it done, there's an important part of the placement being missed.

🔍 Does the paragraph actually relate to the client's topic?

Does the copy naturally lead into the link?

Does the link make sense exactly where it's been placed?

A link sitting beside unrelated content can look paid to an editor, and increasingly, to systems examining context.

Metrics can tell you plenty about the site.

They can't tell you whether the paragraph earned the link.

What does your editorial QC actually check before sign-off?

09/09/2026

There's a problem in SEO: too many people benefit from making it look like magic.

The more mysterious the process appears, the easier it is to position yourself as the magician with the secret answer.

But SEO isn't magic.

🎬 It's more like watching special effects in a movie.

From the audience, something impossible appears to be happening.

Behind the scenes, there are systems, processes, and people who understand how the pieces work.

AI-powered search is no different.

Search engines use hundreds of factors, but the underlying goals are much easier to understand.

Can you demonstrate genuine expertise through your content?

Can the wider web provide enough authority and reassurance through links, mentions, and other signals?

Understand those fundamentals and you can stop chasing tricks and start building something designed to last.

09/09/2026

Can you add schema and suddenly become more visible in AI answers?

Not necessarily.

A study earlier this year found that adding schema on its own didn't materially increase AI citations for pages that were already being cited.

That result makes more sense when you look at it through an entity model.

Schema helps reduce ambiguity. It gives machines structured information that can help clarify what something is.

But recognition is a different problem.

🔎 Recognition develops through a broader set of signals, including consistent descriptions across multiple sources and branded demand.

In other words:

Schema can help clarify an entity that machines already have signals about.

It doesn't manufacture recognition by itself.

🎥 See why this distinction matters:
https://youtu.be/DESQwJFkoCc

09/09/2026

When Google or an AI system tries to understand your business, your website isn't the only source available.

Think of the picture as coming from three places:

1. What your own website says
2. What third parties say about you
3. The patterns that repeatedly appear around your brand across the wider web

🧩 Together, those signals help answer a much bigger question than “What does this page say?”

The system is trying to understand:

“What is this business, and is that answer consistent everywhere it appears?”

That's why optimizing your website alone only addresses part of the problem.

Your wider presence needs to reinforce the same picture.

🎥 We explain the full approach here:
https://youtu.be/DESQwJFkoCc

09/09/2026

Your client opens the monthly report.

They scan two pages.

Then they stop.

If the numbers are there but the story isn't, you've handed them work they now have to interpret themselves, and probably defend to their board.

That's the gap a lot of white-label partners underestimate. 📊

OpenAI reported last week that GPT-5.6 Sol produces 68% fewer factually incorrect responses than the previous model on financial, medical and legal prompts.

The wider lesson is worth paying attention to.

As the tools clients use become more disciplined around facts, sources and dates, the reports we hand them should meet the same standard.

A simple structure helps:

💡 Claim first
🔍 Source next
📝 Assumption stated
🎯 Recommendation last

It's not more work.

It's the same work, ordered so a busy client can understand what happened, why it matters and what to do next.

Good reporting shouldn't make the client work harder to understand the work you've already done.

What's the first thing your clients tend to flag when a report lands?

08/09/2026

Here's something uncomfortable:

Your website is not your business.

Not to Google. Not to ChatGPT.

These systems can use signals from across the wider web to understand what your company is, what it does, and where it belongs.

Yet most businesses spend the majority of their effort managing just one source: their own website.

🧩 That's only part of the picture.

As AI-powered search creates more ways for people to discover businesses, understanding how your brand is represented beyond your own site becomes increasingly important.

Your website tells machines who you say you are.

The wider web helps tell them whether that story holds together.

🎥 See what this means for your strategy:
https://youtu.be/DESQwJFkoCc

08/09/2026

There's an important difference between being cited by AI and being recommended by it.

Being cited means an AI system has found your content useful enough to reference.

Being recommended means it actually names your business when someone asks for a company, service, or solution.

🎯 That's a very different outcome.

If someone receives an answer containing information from your website, they may never visit you.

If the answer says your business is one of the brands they should consider, you're now part of the buying conversation.

That's why citations shouldn't be the finish line for AI visibility.

The bigger goal is becoming a brand these systems understand well enough to recommend by name.

🎥 See how entity clarity helps you get there:
https://youtu.be/DESQwJFkoCc

08/09/2026

Running digital PR for five clients out of one inbox is how cross-contamination starts.

Same journalist pitched twice in a week.

Same data angle recycled.

Same commentary attributed to two competing brands.

It can happen quietly, but the damage isn't quiet. Editors notice patterns, clients lose differentiation, and every account involved becomes a little less distinct. ⚠️

The fix isn't more software.

It's separation by design.

📩 Different sender identities
🔍 Different research pools
💡 Different story angles for each client
🚫 One shared blocklist so nobody accidentally pitches the same outlet twice for competing accounts

There's an AI visibility dimension to this too.

Semrush's 2026 AI Visibility Index found that brand mentions and cited domains can diverge by up to 70% on Gemini.

That matters when you're delivering for multiple clients in the same category.

Distinct sources help build distinct authority.

Merged workflows risk creating merged footprints.

And when two clients are paying you to help them stand apart, that's exactly what you don't want.

What separation rules do you use when two clients compete for the same audience?