Why Scottish Businesses Must Shift From Chasing Rankings to Earning References in AI Search

Scottish firms were alerted in January that generative AI was changing how people discover businesses online. The next problem is tougher: even when demand exists, the customer may never reach their website.

In January, this column stated that generative AI was starting to change how Scottish businesses get found online. That warning still holds. The risk has changed shape.

Scotland has an estimated 384,280 private sector businesses. 99.4% are SMEs. Official Scottish data also shows that 30.7% of businesses were already using some form of AI in December 2025. For most of the Scottish business base, this is now a commercial issue, not a speculative one.

New analysis from Accuracast points to the next stage.

The issue is no longer only whether a business appears in search. It is whether its expertise is clear enough to be surfaced, summarised, and cited when AI sits between the user and the website.

Accuracast has been tracking that shift through its work on generative AI search optimisation.

1. The risk has moved on

The concern voiced in January was simple: businesses could become harder to find.

The next challenge is that visibility and traffic are starting to split. A business can shape the answer and still lose the click (user visit to the website) if the query is resolved without the user needing to click.

Reuters reported in January that UK publishers were already raising concerns about falling click-through rates from AI-generated search features, while the CMA noted that Google still handles more than 90% of general search queries in the UK.

That matters more in Scotland because the economy is dominated by smaller firms. Most do not have the cushion of national brand recognition or large paid media budgets. They rely on being found on search.

2. This is now an operating issue

Scottish firms are already using AI. AI already affects how Scottish firms compete.

The problem is that many businesses still measure search on old terms: rankings, impressions, traditional search clicks. Those metrics still matter, but they miss what AI search changes. A firm might be cited in the answer, mentioned in passing, or bypassed entirely. Those outcomes are not commercially equal.

Google’s own guidance reflects that shift. In its May 2025 guidance on AI search, the company said users are asking longer, more specific questions and follow-up questions inside AI experiences, and advised site owners to look beyond raw clicks to the wider value of visits from search.

3. The regulator has made it a market question

The clearest sign that this has gone beyond a marketing trend came from the Competition and Markets Authority on 28 January 2026. The CMA said Google’s approach to ranking search results, including AI-generated results, should be fair and transparent for businesses. It also proposed stronger attribution for publisher content in AI features and better complaints processes where ranking changes have significant effects on UK businesses.

This is no longer only about whether marketers like AI Overviews. It is about how visibility is allocated when the interface itself starts answering the query. Reuters’ coverage of the matter noted the pressure from publishers to be able to opt out of AI Overviews without harming their general search position.

4. From rankings to references

Google’s public guidance still says the fundamentals matter: unique content, strong page experience, crawlability, indexable content, and structured data that matches what users can actually see on the page. But the search journey is changing. AI search is built around resolving tasks, not just returning pages.

The contest is no longer only about where a page ranks. It is about whether a business becomes a source that an AI system can extract, trust, and cite. These AI platforms reward clarity, specificity, and real expertise. They punish thin service pages and generic copy.

That is where Accuracast’s methodology is useful. Its work on AI search optimisation is built around a simple point: visibility in AI-led search depends on whether a brand’s content is structured and authoritative enough to be used in conversational systems such as ChatGPT and Google’s AI Overviews. For Scottish businesses, ranking is now only part of the contest. Mentions and citations matter too.

5. What Scottish businesses are still getting wrong

Many firms still treat this as a technical SEO issue. It is a business visibility issue.

Scottish Enterprise’s advice on AI adoption is simple. Start with strategy and tie any AI work to commercial outcomes such as revenue, lower costs, better customer experience, or faster decisions.

The same applies here. If a business cannot identify which pages drive enquiries, which topics attract qualified visitors, or which proof points make it credible, it is not ready for AI search.

The mistakes are basic, but costly:

• Key information buried in PDFs or slide decks

• Thin service pages with little depth or proof

• Weak internal linking to core commercial pages

• Structured data that does not match the visible copy

• Local pages that mention a place but say little about the service

• Useful expertise locked behind forms, image-heavy pages, or files search engines struggle to read

Google’s guidance is clear. Pages need to be crawlable, indexable, and useful, with content that can be understood and accessible within normal page text if they are to perform well in AI search features.

For Scottish SMEs, that is the gap. Many know AI matters. Few have improved the pages on their websites that actually need to earn trust.

6. What to fix in the next 90 days

The right response is not to chase easy wins and tricks. It is to improve the parts of a site that AI search still relies on.

Make core pages eligible

Google says a page must be indexed and eligible to show a snippet in Search if it is to appear as a supporting link in AI Overviews or AI Mode. If the page is blocked, thin (light on original content), or missing indexable text, it is already at a disadvantage.

Rewrite weak service pages

AI search rewards pages that answer specific questions clearly. Google says users in AI search ask longer and more detailed queries, often with follow-up questions. Thin brochure copy is less useful in that environment.

Focus on pages that drive revenue:

• Core services

• Sector pages

• Location pages

• High-intent comparison or explainer content

Strengthen internal links and on-page copy

Google still relies on the basics: internal linking, visible text, structured data that matches the page, and good page experience.

Check:

• Whether your most important pages are linked from other relevant pages

• Whether headings and copy state what the page actually covers

• Whether schema reflects the visible content

• Whether business details are current across search surfaces

Control what AI can preview

Google allows site owners to manage how content appears using tools such as no snippet, data-nosnippet, max-snippet, and no index. That matters for proprietary research, gated assets, or pages where you want visibility without giving away too much in the preview.

Measure the right outcomes

Clicks still matter. They are no longer enough.

Google’s guidance on AI search points site owners toward broader performance signals, including conversions and visitor value, not just raw click totals.

Track:

• Qualified inquiries

• Assisted conversions

• Branded search growth

• Lead quality from organic traffic

• Visits to high-intent pages

Use the support already on offer

Scotland already has support in place. The Scottish Government launched a national AI adoption programme for SMEs, and the programme offers courses, consultancy support, roadshows, and grant support for eligible businesses. Scottish Enterprise is also pushing firms to take a more strategic approach to AI adoption.

7. Consumer behaviour is moving faster than many firms think

User behaviour is already moving.

Ofcom said in December that Google Search is still used by 82% of UK adults, but AI-generated search summaries are now a routine part of that experience for many users. In the same report, Ofcom said ChatGPT recorded 1.8 billion UK visits in the first eight months of 2025, up from 368 million in the same period of 2024.

Which? found that 51% of UK adults use AI search tools in their personal lives to look for products, services, and advice online. Among 18 to 34-year-olds, that rises to 75%. It also found that 24% of UK adults use AI search tools frequently.

Which? also found that 55% of AI search users ask follow-up questions to improve results, while 23% often or always check one AI tool against another.

That matters for Scottish businesses because a weak page can still win the odd click. It is much harder for weak content to become a trusted citation in a conversational search journey.

8. The opportunity is still there

This shift will not reward the biggest budget every time. It will reward the clearest source.

Scotland is already pushing SME AI adoption. The Scottish Government said in September 2025 that a national AI adoption programme for SMEs would form part of its wider AI Scotland programme.

The programme now offers practical support for Scottish SMEs that want to start using AI technologies.

Smaller Scottish firms still have room to gain. They do not need to outspend national competitors to improve their position in AI search.

They need stronger commercial pages, better proof, cleaner site structure, and content that answers real questions more clearly than the generic copy still sitting on many websites.

Google’s own guidance still points back to useful, original content and sound technical foundations.

The old search model rewarded ranking.

The new one rewards being useful and trustworthy enough to cite or mention.