I haven’t written a proper blog post here for a while.
Classic developer behaviour, really. Spend years telling clients they should publish useful content, then quietly let your own blog gather dust like a gym membership in February.
In my defence, the web has been doing its usual trick of changing shape while everyone is still trying to get their websites into the shape we all agreed was sensible five minutes ago. One minute we’re arguing about Core Web Vitals, then everyone is using ChatGPT to write LinkedIn posts about how ChatGPT is going to replace the people writing LinkedIn posts.
But under the noise there is something real happening.
Search is changing.
Not “Google changed the colour of a button and now three SEO consultants have made a carousel about it” changing. Proper changing. The way people find answers, compare suppliers, judge expertise, and decide who to trust is moving from ten blue links towards AI summaries, conversational search, answer engines, agents, citations, community snippets, and whatever acronym gets invented before I finish this sentence.
The term floating around for this is usually GEO: Generative Engine Optimisation. Sometimes AEO: Answer Engine Optimisation. Sometimes AI visibility. Sometimes “new SEO”, which is a phrase that makes me want to walk into the sea but is also annoyingly understandable.
The useful version is simpler:
Can AI systems understand, trust and recommend your business when someone asks a relevant question?
That’s the bit I care about.
Not because I’m desperate to sell another acronym. We have enough of those. I care because the work underneath it is exactly where a lot of businesses are weak: unclear positioning, thin content, messy websites, slow pages, invisible expertise, no structured data, disconnected CRMs, no useful follow-up, and a “Contact us” page doing the emotional labour of an entire sales team.
AI hasn’t made that less important.
It has made it more obvious.
The website is not dead. Sorry.
Every few years someone announces that websites are dead.
Social media was going to kill websites. Apps were going to kill websites. Voice search was going to kill websites. TikTok was going to kill Google. QR codes somehow survived both a pandemic and several crimes against restaurant menus.
And now AI is apparently going to kill websites.
I don’t buy it.
What I do buy is that the role of a website has changed.
Your website is no longer just a brochure people land on after Googling your name. It is increasingly the source material used by humans, search engines, AI assistants, answer engines and agents to work out:
- who you are
- what you actually do
- whether you know what you’re talking about
- whether your business is alive or just technically online
- whether you are worth quoting, citing, comparing or recommending
That last bit matters.
AI search is not just ranking pages. It is summarising answers. It is comparing options. It is pulling in sources. It is giving people a “good enough” answer before they ever see your carefully designed hero section with the lovely gradient I definitely spent too long nudging by 3 pixels.
So yes, your website still matters.
It just has a new job interview to pass.
GEO is not magic. It is mostly clarity.
The slightly annoying thing about GEO is that the name makes it sound like there is a secret switch somewhere.
There isn’t.
For most businesses, GEO starts with ordinary good digital work:
- clear pages that explain services properly
- original expertise instead of beige filler
- useful case studies
- proper author and business signals
- structured data and schema
- fast pages
- content that answers real buying questions
- a site architecture that does not resemble a drawer full of old charging cables
- analytics, forms and CRM follow-up that actually work
The difference is who you are writing and building for.
Before, you were mainly trying to satisfy a search engine and a human.
Now you are also trying to satisfy an AI system that is reading across many sources, compressing the answer, and deciding whether your business deserves to appear in the summary, citation, comparison, or “you should probably speak to this person” pile.
That sounds a bit dramatic, but it’s already happening. Google has AI Mode in the UK. ChatGPT has search with cited web sources. Google is actively adding more links, source previews and firsthand perspectives into AI Mode and AI Overviews because even Google knows a dead-end answer box is not enough.
The web is still in the loop.
But it is a different loop.
The problem with most “AI strategy”
Here is where I get a bit twitchy.
There is a huge amount of AI advice around at the moment that sounds impressive until you ask the second question.
“You need an AI transformation roadmap.”
Fine. What are we transforming?
“Your business processes.”
Which ones?
“Key workflows.”
Which workflows?
“We’ll do a discovery workshop.”
At this point I start gently checking whether there is an exit behind the fern.
The useful AI work is not usually “let’s bolt a chatbot onto the website and hope the future is impressed.”
It is things like:
- your sales enquiry process is slow and inconsistent
- your quote follow-up depends on one person remembering things
- your website does not explain the real value of the service
- your team is manually copying data between tools like it’s 2009
- your best expertise is trapped in your head, emails, PDFs or WhatsApp messages
- your content is not structured well enough for search engines or AI systems to trust it
- your CRM technically exists but everyone treats it like a haunted filing cabinet
That’s where AI becomes useful.
Not as the shiny object. As the tool that helps fix the actual business problem.
The awkward truth: you still need someone who can build
This is where I think a lot of the AI conversation gets silly.
There are people who can talk about AI all day. Some of them are very good. Some of them have discovered the word “agentic” and are now legally required to use it every twelve seconds.
There are also people who can build websites, tools, automations, integrations and admin systems.
The useful overlap is smaller.
Because advising a business on AI is not just about knowing the tools. It is about knowing what should exist, what should not exist, what will break, what users will ignore, what data is too messy, what the client actually means, and how to turn the idea into something people can use without needing a 42-minute Loom video.
I’ve spent years doing the “boring” bits: WordPress builds, bespoke themes, plugin fixes, WooCommerce oddities, API integrations, form logic, speed work, hosting, weird Safari bugs, ACF deciding to escape HTML and make a perfectly normal Tuesday feel like a hostage negotiation.
That stuff is not separate from AI.
It is the foundation that makes AI useful.
If your website is slow, vague and badly structured, AI does not magically make it strategic. It just gives you a faster route to a bad answer.
If your CRM is a mess, AI does not create operational excellence. It creates a mess with nicer sentence structure.
If your offer is unclear, AI will not fix the positioning. It will generate seven versions of the same unclear thing and one of them will have an emoji in it.
What I’d actually check first
If a business asked me tomorrow, “are we ready for AI search?” I would not start with a prompt library.
I’d start here:
1. Can a stranger understand what you do in 10 seconds?
Not your mate. Not your existing client. A cold visitor with six tabs open and a cup of coffee going cold.
If the answer is no, AI systems probably won’t understand you cleanly either.
2. Do your important services have proper pages?
One “Services” page with six vague boxes is not enough anymore.
If you want to be understood for AI consultancy, automation, GEO, WordPress, web apps, or anything else, each area needs enough depth, examples and context to stand on its own.
3. Is your expertise visible?
AI search leans heavily on signals of real experience.
That means case studies, opinions, author context, original examples, explainers, comparisons, and proof you have actually done the work.
Not “we leverage cutting-edge solutions to empower digital transformation.”
That sentence should be taken outside and hosed down.
4. Is your site technically readable?
Semantic HTML. Schema. Fast pages. Clear headings. Sensible internal links. No 19MB hero video because someone once said “cinematic”.
The boring stuff still matters.
It always does. That’s why it’s boring.
5. What happens after the click?
This is the bit people miss.
Getting found by AI search is not the win. It is the start.
If someone clicks through, asks a question, fills a form, books a call, requests a quote, or downloads something, what happens next?
Does it go into a CRM? Does it trigger the right follow-up? Does the client get a useful next step? Does anyone know where the lead came from? Or does it vanish into a shared inbox called “info” that nobody wants to own?
AI visibility without a working sales process is just a more modern way to leak enquiries.
So what am I doing with this?
I’m being fairly deliberate about where I’m taking Guwii.
I’m still building websites. Good ones. Bespoke ones. WordPress ones. Web apps. Tools. Systems. The work still needs design, code, UX, speed, security, hosting, content and taste.
But the lens is changing.
More of my work now sits around:
- AI strategy for real businesses
- GEO and AI visibility
- AI-ready websites and content structures
- custom internal tools
- workflow and sales automation
- CRM and API integrations
- practical AI assistants
- monitoring and technical intelligence through Scavo
In other words: not “AI for the sake of AI”.
AI where it helps a business move faster, explain itself better, reduce manual work, or make better decisions.
That’s the bit I think is going to matter, whether the client is a small business today or a larger organisation tomorrow.
Because the valuable person in the room won’t be the one who has memorised the most tools.
It will be the one who can say:
“That’s worth automating.”
“That needs a better user journey first.”
“That should be a structured content system.”
“That is a terrible idea, but if you absolutely insist, let’s at least make it secure.”
“No, your chatbot should not answer refund questions by hallucinating a returns policy like a tiny legal liability with a smiley face.”
That is the work.
In conclusion
AI search is not the end of websites.
It is the end of websites that can get away with being vague.
Your site, content, data, tools and workflows are now being read by machines before, during and after humans interact with you. That doesn’t mean every business needs to panic-build an AI chatbot by Friday.
It does mean the digital foundations need to be cleaner, clearer and more useful.
GEO is not a magic spell.
It is a reminder that if you want to be recommended, cited, compared or trusted in an AI-shaped web, your business needs to give the machines and the humans something solid to work with.
If you’re looking at your site, your content, your sales process or your internal tools and thinking “this is probably not ready for that”, you’re probably right.
And yes, I can help with that.
No 80-page transformation deck required. Unless you specifically want one, in which case I’ll at least try to make it less painful than the average airport PowerPoint.