TL;DR Short answer: most small local businesses in our region do not need a full “agent‑ready” rebuild right now – but a few low‑effort technical tweaks and better basic signals will put you in a good position as AI search rolls out. This post explains who genuinely needs to act now, who should quietly plan for it, and who can safely wait, with simple next steps for each.

That sounds like a natural question a business owner in your patch might ask, and it makes clear this is about timing and priority, not jargon.

 Over the last few months, terms like “agentic search”, Universal Cart, llms.txt and “agent‑ready websites” have started popping up in SEO blogs, socials, webinars and tools. If you run a business in Cumbria, Dumfries & Galloway or the wider Borderlands, the honest question is probably simpler: do we actually need to do anything about this now, or can it wait until the dust settles?

 This article answers that question as directly as possible. It pulls together three strands – Google’s I/O 2026 announcements about the new AI‑powered search experience, independent analysis on how AI search is changing visibility and brand preference, and Suganthan Mohanadasan’s very practical framework for the “minimum viable” agent‑ready site – and then applies them to the kinds of small businesses and ecommerce firms we actually have in this region.

 The goal is not to sell you another AI product. It is to give you a clear sense of whether you are in the “act now”, “plan for it” or “can safely wait” bucket – and, if you do need to act, what the smallest sensible next steps look like for your website.

What does “AI search” even look like?  AI has been an increasingly present part of search technology for years; but what has changed, is the new AI / Large Language Models that have appeared on the scene and which are accessible to domestic and business users at relatively low cost or for free. Initially they were answer machines but they have increasingly become a search/answer engine hybrid that also powers everyday processes behind the scenes of businesses and other systems. This is steadily interrupting people’s old search habits and has forced Google’s hand into deploying its own AI more rapidly than it had initially planned to.

The graphic below gives a very basic glimpse into how the new AI powered search tech reads, steals and regurgitates content as answers if it finds your website / brand worth mentioning.

A useful way to think about this is to split websites into three broad groups:

1. act now,
2. plan for adoption soon,
3. optional/ skip it completely for now.

Framing it that way may help your businesses avoid treating every new “AI protocol” post on Linked In as urgent when, in reality, adoption is uneven and the returns depends heavily on the audience your business needs, the type of website you operate and what your business model is.

1. Act Now – Because it will probably be a good investment

The “act now” group includes technical SEO teams and agencies, brands who are already tracking AI search citations or their share of voice, companies which are producing AI-first products such as chatbots and AI APIs, and businesses selling to audiences that already run AI agents on the web. For these organisations, you are likely to have the resources and budget to experiment; and early adopter advantage is likely to be strong for you because your audience, product or measurement model already overlaps with AI agent systems to some extent.

2. Plan For It – Because you will be affected soon

The “plan for it” category is where many growing SMEs sit: mid-size content publishers, ecommerce sites with structured product data, local businesses serving high-intent users where competition is strong, and brands whose competitors are starting to deploy agent protocols: think – booking systems in the Lake District, for instance. Your businesses probably don’t need a full rebuild of website and systems to benefit from AI search just yet, but you can see there is going to be a commercial benefit from understanding the direction of travel and making a small set of low-risk improvements over the next 6 to 12 months. You will have already noticed a decline in organic website traffic and clicks even though your website has strong ranking against competitors. (Yes – Google is stealing your visitor traffic).
My own judgement at present is that most Cumbrian businesses who care about their digital visibility fall into this category. Whether you do flooring, paralegal services, business tech, dentistry, accountancy, business support services, legal work, retail or run a venue then the move to AI based search is going to affect you as your customers adjust their search habits to the new developments .

3. Optional or Skip It

Personal blogs, hobby sites, small brochure sites with very little traffic, and organisations whose audiences are largely offline or non-technical are in the optional or skip it completely categories for now. For these sites, the work is not necessarily wasted, but the value is unlikely to build up quickly enough to justify urgency.

One of the most useful conclusions from Mohanadasan’s article (endorsed by Lily Ray – one of the more influential and down to earth SEO/digital marketing practitioners I follow) is that only a handful of components currently offer a sensible return for effort put in now. These are the items which do have some clear precedent and where some kind of consensus is growing:

• txt controls for AI bots
• sitemap discovery
• semantic HTML
• accessibility
• llms.txt 

As you will see these priorities for AI search… are also basic SEO tactics apart from the one specialised tactic specific to AI engines.

1. AI bot rules in robots.txt: a fast, low-effort way to control which compliant bots may crawl the site, read its content, use the content in AI answers, use it as training material.
2. Sitemaps: still a basic for discoverability and clarity by search engines and AI crawlers.
3. Semantic HTML: Quality HTML that clearly describes what each part of the page is and does, improving basic SEO while also making the page easier for AI agents and assistive tech to understand. It’s useful regardless of AI because this improve human usability and machine interpretation at the same time.
4. Accessibility: A site that is genuinely accessible for people of all abilities – with clear structure, labels and keyboard navigation – is also far easier for AI agents and search engines to read and understand accurately, improving your chances of being mentioned in an answer.
5. llms.txt: a short machine-readable summary of important content and URLs; not a guarantee of citations, but quick and cheap to publish. This is increasingly recognised by tools such as Chrome Lighthouse (in almost direct contradiction of Google’s published information about optimising for AI visibility… by a separate Google team!)

For most relatively small local service firms, the first four items are the realistic priority list.

At which point many of you will say. “But that’s what we already do for our SEO?”… and you aren’t wrong!

There are more advanced components that may come into play, such as commerce protocols, Agent Skills, A2A cards and API catalogues. For most business websites though these are not part of their operation.

It is vitally important to understand that Google still dominates the bulk of search behaviour. Most recent estimates putting its global share at roughly nine in ten searches, even with AI tools and alternative discovery habits slowly chipping away at that dominance. At the same time, search behaviour had already been shifting before the recent AI wave, with many users, especially younger ones, starting local discovery journeys on TikTok, Instagram, YouTube and other social platforms for reviews, tutorials and local recommendations.

That context matters because Google’s own I/O 2026 announcements show they are making a brazen attempt to keep those discovery and decision journeys inside Google itself, by expanding their own AI Mode, the new – soon to be everywhere -Intelligent Search Box, shopping flows and AI agents taking action for you.

In practical terms, the web’s classic blue links are not disappearing, but your links are being pushed further from the top of the default search experience; as AI answers, recommendations and actions take up more and more space at the top of the results pages as Google tries to monopolise a web user’s first interaction. You will see your visitor clicks and traffic reducing as Google effectively steals that traffic and feeds it your information or content without a visit to your website ever happening.

For local service firms, that means the immediate priority is not building the most advanced “agent-ready” website technology stack possible. The bigger win is still getting the fundamentals right. Yes those same SEO fundamentals I’ve been banging on about for the past 15 years! Clear service pages, accurate Google Business Profiles, strong local trust signals, structured data where appropriate, and encouraging specific customer reviews that describe the job, location and outcome or benefit of what you did for them.

None of this is silver bullet material. Publishing llms.txt, refining robots.txt or adding newer agent protocols does not guarantee AI citations, referral traffic or increased prominence in AI search results.

The technology, the product and services that people are inventing to capitalise on business owners’ discomfort with these changes and the standards that are emerging in this area are changing quickly. Adoption is and is likely to remain uneven. Some of the more experimental protocols may fragment or be replaced before they become mainstream. That is why the minimum viable list is so useful: it focuses on practical steps with low downside and avoids overcommitting to unstable standards too early.

For most small business sites, the sensible course is straightforward: keep up your SEO work, implement the low-risk technical basics that appear to be useful for AI search, improve the visibility and trust signals that already matter in classic search, and monitor how AI search actually develops in the UK market before investing in anything more elaborate.

That means treating agent-readiness as an extension of solid technical SEO and content clarity, not as a separate panic project.

For ecommerce firms, especially those with niche products, rich specifications, the recommendation is slightly different: make the site agent-ready by default, at least at the minimum viable level. That keeps options open for future shopping integrations and reduces the risk of needing another technical overhaul when agentic commerce becomes more prevalent.

Sources and Further Reading

This article draws on a mix of official product announcements, industry reporting and independent analysis to separate immediate practical actions from longer-term speculation.

• Suganthan Mohanadasan on the minimum viable agent-ready site
  The practical framework used here — including the “act now / plan / optional / skip” categories and the minimum viable list of technical actions — is based on Suganthan Mohanadasan’s guide to the minimum viable agent-ready site, as excerpted in the supplied notes. See:
  How to Make Your Website Agent-Ready.
• Google’s own I/O 2026 search announcements
  Product-level context on AI Mode, the Intelligent Search Box, search agents and Google’s broader AI search direction comes from Google’s I/O 2026 materials and official blog coverage, for example:
  Google Search’s I/O 2026 updates: AI agents and more.
• Coverage of Universal Cart and agentic commerce
  The sections on shopping, structured product data and the likely implications for ecommerce are informed by reporting on Google’s Universal Cart and the broader agentic commerce push, including:
  Google’s frictionless retail vision starts with “universal cart”
  and
  What Google’s Universal Cart Means For Agentic Shopping.
• llms.txt and agent-friendly technical signals
  References to llms.txt, Chrome Lighthouse and related “agent-readiness” ideas build on the supplied technical notes and wider discussion around AI crawler discoverability, for example:
  Google adds llms.txt check to Chrome Lighthouse.
• AI search, entity building and reputation
  The broader interpretation of AI search, entity clarity, reputation and small-business visibility draws on the comparative analysis compiled from Claude, Gemini and Andrew Holland’s writing in the supplied materials.
• Search behaviour beyond Google
  The argument that Google still dominates traditional search while social and AI platforms are gradually chipping away at attention and discovery behaviour is informed by recent reporting and market-share analysis, including:
  Search Engine Statistics 2026 and
  Gen Z Dumping Google For TikTok, Instagram As Social Search Wins.