We’re at BrightonSEO this week, and in our training session on AI search we were looking at one very specific question:
What does good copywriting look like in a world of AI search, AI Overviews and large language models?
Not in theory. In practice.
Because while there is a lot of noise around AI and SEO, the reality is much simpler and much more immediate. Search behaviour has already changed. Google is already answering more queries directly. And AI tools are already deciding which content gets surfaced, summarised or ignored.
What Do The Numbers Say in AI Search?
Some of the numbers shared during the session made that clear:
- 60% of Google searches now end without a click
- In AI Mode, that rises to around 93%
- And yet, Google still sends significantly more traffic than any AI platform
So, this is not about abandoning SEO. It is about understanding how the role of content is changing.
We are no longer writing just to rank. We are writing to be selected, understood and used.
Google still sends vastly more referral traffic than any AI platform. In fact, it is still responsible for around 190 times more traffic than ChatGPT.
So, this is not about abandoning SEO. It is about understanding how it is evolving.
People are still searching. They are just getting answers in different ways.
Which means your content now has a different role to play. It needs to be clear, credible, and structured enough to be selected, summarised, and cited.
How Google Builds an AI Overview
One of the most useful frameworks from the day broke down how AI Overviews are actually created:
- Query analysis and fan-out
- Source retrieval
- Synthesis
- Citation selection
- Placement
The important shift here is in the middle.
Google is no longer choosing a single “best” page. It is building an answer from multiple sources.
There is no longer one winner. There is a collection of contributors.
That changes the job of copywriting quite significantly.
Three Ways Your Content Can Appear in AI Overviews
Another simple but important takeaway was around how your content can show up in AI-generated answers. There are three ways:
- Citations: your page is linked as a source
- Brand mentions: your name appears without a link
- Contributions: your content is used, but you are not credited
That last one is uncomfortable, but it is part of the reality of how these systems work.
It also reframes success. Visibility is not just about traffic anymore. It is about influence within the answer itself.
Links Still Matter But Not in The Same Way
We saw familiar data around links and rankings, such as how pages with more links still tend to perform better.
But the role of links has shifted slightly.
They are not just helping you rank. They are helping establish you as a credible source worth including in an answer.
There were a few supporting patterns worth noting:
- High-authority domains are cited more frequently
- Content already performing well in search is more likely to be reused
- Recently updated content is significantly more likely to be cited
In fact, blogs that appeared in AI answers were very often updated within the last 12 months.
Freshness now has a very practical benefit.
Where LLMs are Actually Getting Their Information From
This is where things become more interesting from a copywriting perspective.
The most cited sources across AI systems are not just websites. They include:
- Wikipedia
- YouTube
- Established publishers
This explains why one of the clearest recommendations from the session was to encourage clients, particularly in B2B, to build a presence on LinkedIn.
Your blog is no longer the only place your authority is built. The wider ecosystem matters just as much.
From Deterministic to Probabilistic Search
A useful way to understand the shift is to compare traditional search with how LLMs work.
Traditional search engines:
- Crawl and index pages
- Match queries to keywords
- Rank results in order
LLMs and AI search:
- Analyse patterns across large volumes of content
- Run multiple queries at once
- Combine findings into a single response
- Generate an answer rather than rank one
It is less about retrieval and more about construction. This means your content needs to be easy to extract from, interpret and reuse.
The Queries That Trigger AI Overviews
Not every search leads to an AI Overview. There are clear patterns.
Queries that are more likely to trigger AI responses:
- Informational and how-to queries
- Complex or multi-part questions
- Definitions and explanations
- Comparison queries
Less likely:
- Branded searches
- Breaking news
- Local queries
- Purely transactional intent
This gives a clear direction for content strategy, as many people and companies these days want to focus on the questions AI is most likely to answer.
Writing Content That AI Can Actually Use
This is where copywriting becomes more intentional. A few practical rules stood out:
- Front-load your answers
Do not build slowly towards a point. Start with it. AI systems tend to prioritise content that gets to the answer quickly. - Keep sentences simple
Clear subject-verb-object structures are easier to process. Readability tools are no longer optional. - Use atomic facts
Keep statements short, specific and verifiable. Ideally under 20 words. Avoid vague claims. - Be explicit with entities
Name brands, tools, people and places clearly. Specific nouns help machines understand context. - Write straightforward headlines
Your headline should say exactly what the page does. Clarity beats cleverness.
Structure Matters More Than Ever in Copywriting
People scan before they read. AI scans even faster. A few practical formatting we learned from the session:
- Aim for 120–180 words between headings
- Write subheadings that tell a clear story on their own
- Use jump links on longer pieces
- Plan your headings before writing the body
- Structure content logically, with clear hierarchy
If someone only read your headings, they should still understand the content.
Topics Matter, Not Just Keywords
Now, this is not a new idea, but it has become much more important.
We do not need pages filled with keywords. We need pages that are clearly about something.
That means:
- Building topic clusters
- Creating connected content
- Maintaining a clear site structure
- Using clean URLs and navigation
- Supporting everything with internal linking and canonical tags
All of this helps both search engines and AI systems understand your content more easily.
EEAT and The Ongoing Importance of Trust
Experience, expertise, authoritativeness and trust still underpin everything. But trust is the deciding factor.
Even well-written content will struggle if it does not appear credible.
This is where activity beyond your own site becomes important:
- PR and media mentions
- Thought leadership
- Consistent brand presence across platforms
AI systems are not just evaluating your page. They are evaluating your reputation.
Small Details That Make a Difference in AI Overviews
Some of the most useful takeaways were small but practical:
- Recently updated content can receive significantly more citations
- Repurposing blog content into LinkedIn posts helps reinforce authority
- Encouraging third-party discussion increases visibility
- Auditing where your brand is mentioned can highlight gaps
These are all manageable changes, but they have a real impact.
What Good Content Looks Like in 2026
If you strip everything back, content that performs well in this environment tends to have the same characteristics:
- A clear answer early on
- Logical structure
- Specific, verifiable information
- Credible authorship
- Simple, readable language
- Regular updates
It is not about doing something completely new. It is about doing the fundamentals well, consistently.
Final Thoughts
The biggest shift is not technical. It is a mindset change. We are no longer optimising purely for clicks, we are optimising for inclusion, interpretation and citation.
Traffic still matters. But it is no longer the only measure of success.
The real question is whether your content is part of the answer. Contact us today to learn how we can help with your online presence through our bespoke organic search services.
