The SEO industry has undergone many changes over the years, but the impact of AI is different and it is more dramatic than anything that came before. It is a structural change to what search actually is and a tool for how businesses operate. And most SEO agencies, (including mine for a while) have been slow to say that out loud.
Keep reading if you would like to know how AI is changing SEO, why I resisted it for so long, and what changed my mind.
What SEO Involves
Before I get into the impact of AI on SEO, I want to be clear about what SEO actually is. A properly run SEO campaign covers at least six distinct areas. Each one requires expertise, time, and consistent execution.
| Pillar | What it involves | How AI is changing it |
|---|---|---|
| Technical Audit | Crawling for broken links, duplicate content, redirect chains, slow pages, rendering issues | AI agents can now interpret crawl data, rank issues by impact, and draft fixes, cutting days of specialist work to hours |
| Keyword & Competitor Research | Identifying search intent, gap analysis, competitor benchmarking | Semantic clustering, intent mapping, and gap analysis across thousands of variants now done in a fraction of the time |
| Technical SEO | Schema markup, canonical tags, XML sitemaps, hreflang, site architecture | Schema, meta tags, and sitemaps can be generated, validated, and maintained by AI agents on custom workflows |
| On-Site Optimisation | Title tags, meta descriptions, heading structure, internal linking, image alt text | AI can audit and generate on-page elements at scale, with human review at the quality checkpoint |
| Content Creation | Researched, optimised editorial copy (on a busy month, over 100,000 words for our clients alone) | AI-assisted production shifts the role from writing to prompting, quality checking, and strategic direction |
| Link Building | Earning backlinks from authoritative sources, still one of Google's strongest trust signals | AI assists with prospecting and outreach drafts; human relationships and editorial judgement remain essential |
That is what you are paying for when you hire an SEO agency like Opace.
How AI Is Changing SEO
Automated Audits and Technical SEO
Technical SEO was always the most automatable part of the job, even before AI. Tools like Screaming Frog, Website Auditor and SurferSEO have been crawling sites and flagging issues for years. What AI changes is the interpretation layer.
A crawler can tell you that a page has a slow server response time or broken links. An AI agent, fine-tuned with the right prompting and trained on your site's specific architecture, can tell you why, rank the issues by likely impact, draft the fix, and push it to a staging environment for review.
Schema markup, meta tags, sitemaps, canonical tags - all of this can now be generated, validated, and maintained by agents running on custom workflows. The technical SEO work that used to take a specialist several days per site can now be handled in hours, with a human reviewing the output rather than producing it from scratch.
Keyword and Competitor Research
Keyword research used to involve pulling data from multiple tools, building spreadsheets, identifying clusters, mapping intent, and then cross-referencing against what competitors were ranking for. A thorough research phase for a new client could take the best part of a week.
AI-assisted research changes the quality ceiling. Pattern recognition across thousands of keyword variants, semantic clustering, intent mapping, and gap analysis against the top ten ranking pages. All of this can now be done in a fraction of the time, and with a level of cross-referencing that would be impractical to do manually. The output still needs a strategist to interpret it and make decisions, but the raw analytical work is no longer the bottleneck it was.
AI vs. Human Copywriters
A skilled SEO copywriter producing a well-researched, properly optimised 4,000-word article takes between one and three days. Factor in the briefing, the research, the draft, the review, the revisions, and the final quality check, and you are looking at a meaningful cost per piece. For a client on a content-heavy SEO programme, those costs add up quickly.
AI can produce a comparable piece for pennies. The human time involved shifts from writing to prompting, quality checking, providing feedback, and iterating. A professional SEO strategist spending two to three hours on a piece, including the research brief, the AI generation, the quality review against tools like SurferSEO, and the final editorial pass, can produce something that matches or exceeds what a human copywriter would produce in three days.
The expert is still essential but the key is less around copywriting and more the knowledge of what to write, how to structure it, what gaps to fill, what the competition is missing, and how to quality-check the output.
To put some numbers around this, here is how the three approaches compare using industry averages:
Link Building and Outreach
AI can help identify prospects, draft outreach emails, and analyse the authority of potential linking domains. AI agents can even automate content submission and publication processes online. But the actual relationship-building, the pitching of genuinely original ideas to editors, and the creation of content worth linking to - this all still requires human input.
That said, AI is now very good at creating the content that earns links, because the speed and scale of production means you can create more genuinely useful resources in less time.
Why I Initially Resisted AI SEO
Lessons Learned and Reason for Caution
If you have been doing SEO long enough to remember the pre-Panda world, you will understand why my default position on AI content was scepticism. In the early days, you could rank almost anything with enough keyword repetition and enough links.
Then Panda arrived in 2011 and wiped out content farms. Penguin followed in 2012 and destroyed link networks. Every time the industry found a shortcut, Google found a way to close it.
Mass-produced content, spun articles, auto-generated pages, etc. have all been targeted at one point or another. So when AI content tools started appearing, my instinct was the same as it had been every other time to assume it was a shortcut, and shortcuts get penalised.
For a long time, I held the line. My team produced human-written, individually optimised content (and still do). On a busy month, we were producing over 100,000 words of copy for clients, all of it written and reviewed by people. It was expensive, it was slow by comparison, but it was safe.
Google's Official Stance on AI Content
Google's official guidance on AI-generated content, published in February 2023, is clear on the distinction that matters:
Any means of manipulating rankings is a violation of their spam policy.
Regardless of who (or what) creates content, Google's ranking systems are designed to reward original, high-quality content that demonstrates E-E-A-T: experience, expertise, authoritativeness, and trustworthiness.
Helpful Content Updates target content that is produced primarily for search engines rather than for people. Google is direct about what this means in practice:
Content should be written for people and not search engines.
A generic, low-value article written by a human is just as likely to be demoted as a generic, low-value article written by AI. What Google rewards is content that demonstrates genuine expertise, provides original insight, and gives the reader something they could not get from the first ten results they already looked at.
That means the question is not "should I use AI for SEO?" It is "am I using AI to produce something genuinely better?"
When Things Changed
The thing that finally changed my mind was watching what was happening with my own websites and seeing what our clients were doing themselves, without even speaking to us.
We had published a blog post on AI and the future of SEO, written partially by AI and humans, deeply researched, and produced the way we had always produced content. Within a month or two, it was generating over 76,000 impressions and sitting at an average position of 8.7 across sixteen different search queries.
The top queries were not the ones we had explicitly optimised for. They were variations: the future of seo with ai, how has seo changed with ai, will seo be replaced by ai. The article was ranking for the full semantic cluster around the topic, not just the primary keyword. That told me Google was rewarding genuine topical depth and first-hand expertise.
In parallel, we started testing AI content on our own sites with proper process, prompting, and quality control. The results were comparable to our human-written content, and in some cases better, because the research and iteration phases were more thorough than we could afford to be manually.
At the same time, something else was happening with our clients. Some of them had started using ChatGPT to produce content in parallel with our own work. In some cases, their AI content was clearly negatively impacting the overall quality signal of their site, diluting the work we had done.
At this point, I decided the market had already decided that AI was part of the picture, and our own evidence was strong enough for us to fully embrace AI as an agency.
Should I Pay for SEO in 2026?
Google's own guidance on E-E-A-T is relevant here, because it explains exactly what is and is not worth paying for. The implication for SEO spend is that you are not paying for content production. You are paying for demonstrated expertise and the trust signals that come with it.
What You Are Actually Paying For Has Changed
AI has made the production side of SEO faster and cheaper but it has not made the expertise side any less important.
If you use ChatGPT to write an article without doing proper keyword research first, you will produce something that covers the topic at a generic level, misses the specific search intent behind the queries you want to rank for, and looks identical to the hundreds of other AI-generated articles on the same topic.
If you use AI with proper research and a brief, a structured prompt that incorporates expertise and perspective, and a quality review against SEO best practice, you will produce something that can rank and provide genuine value to the reader.
The Risk of Doing It Yourself With ChatGPT
The most common mistake I see businesses make is treating AI as a content vending machine. They put in a topic, they get out an article, they publish it, and they wonder why nothing happens.
Doing SEO badly with AI is not cheaper than doing it well with an agency. It is just a different way of wasting money, with the added risk of actively damaging your site's standing with Google.
What an AI SEO Agency Does Differently
What my agency now offers, alongside our traditional SEO services, is a partially automated SEO service using AI, built around expertise rather than volume.
The process starts with detailed keyword and competitor research, identifying not just what to write about but what angle to take, what the existing content is missing, and what a genuinely better piece would look like. The AI is then prompted and trained on specific rules, processes, and quality standards, not just asked to write something. The output is reviewed against SEO best practice, checked for accuracy, and revised until it meets the standard we would apply to human-written copy.
This saves cost and time but the expertise requirement has not gone away. If anything, it has increased, because the gap between a well-run AI SEO process and a poorly run one is larger than the gap between a good human copywriter and a bad one.
New Breeds of SEO
Generative Engine Optimisation (GEO)
Search is a growing proportion of information-seeking behaviour that now happens through ChatGPT, Gemini, Perplexity, and other AI-powered tools that synthesise answers rather than returning a list of links. Being visible in those answers requires a different kind of optimisation.
Generative Engine Optimisation, or GEO, is about structuring your content so that AI systems select it as a source when generating responses. This means writing in clear, declarative sentences that are easy to extract. It means covering topics with enough depth and specificity that the model can use your content to answer follow-up questions, not just the headline query. It means establishing your brand as a trustworthy entity through consistent, accurate information across multiple sources.
The overlap with traditional SEO is significant but there are specific technical steps that help. For example, clear heading hierarchies, FAQ sections that mirror the conversational queries people ask AI tools, and structured data that helps models understand what your business does and who it serves.
How Traditional SEO Compares to AI SEO
The shift is not just about tools. The entire optimisation model has changed. Here is how the two approaches differ across the signals that matter:
| Factor | Traditional SEO (pre-2023) | AI-First SEO (2024 onwards) |
|---|---|---|
| Primary goal | Rank on page 1 of Google | Be cited in AI Overviews, featured snippets, and generative answers |
| Success metric | Organic clicks and keyword rankings | Impressions, AI citations, brand mentions, and zero-click visibility |
| Content format | Keyword-optimised long-form articles | Declarative, question-answering content structured for extraction |
| Key ranking signals | Backlinks, keyword density, page authority | E-E-A-T, entity clarity, topical depth, structured data |
| Technical priority | Crawlability, page speed, mobile-first | All of the above, plus schema, llms.txt, and entity disambiguation |
| Link building | Volume of backlinks from authoritative domains | Same, but content must be genuinely citation-worthy to earn links |
| Tooling | Screaming Frog, Ahrefs, SEMrush, manual copywriting | AI agents, semantic analysis tools, GEO monitoring, LLM visibility tracking |
Answer Engine Optimisation (AEO)
A substantial proportion of Google searches now end without a click. The user gets their answer from the AI Overview, the featured snippet, or the knowledge panel, and they never visit a website.
Answer Engine Optimisation is about remaining visible in that zero-click environment. It means structuring content to answer specific questions directly, in the first paragraph of a section rather than buried three hundred words in. It means using the exact phrasing that people use when they ask questions, because AI systems match on natural language rather than keyword density.
For agencies and businesses, this changes the metrics that matter with impressions and AI citations becoming as important as clicks. Brand visibility in AI-generated answers is a form of reach that does not show up in traditional analytics but absolutely influences purchase decisions.
Large Language Model Optimisation (LLMO)
Beyond individual search queries, there is a longer-term question about whether your brand exists in the knowledge base of the major AI models. When someone asks ChatGPT to recommend an SEO agency in Birmingham, does your agency come up? When someone asks Gemini about the future of SEO, does your content get cited?
This is what Large Language Model Optimisation addresses. It involves ensuring your brand has clear, consistent, machine-readable information available across the web.
None of this replaces the fundamentals of SEO. It builds on them. If Google cannot crawl your site, changes are that AI won't cite it either.
Technical SEO Is Still the Foundation
This is the part that gets lost in conversations about GEO and AEO and the future of search. All of the new optimisation strategies depend on the same foundation that traditional SEO has always depended on.
If your pages are slow to load, your JavaScript rendering is blocking Googlebot from seeing your content, your internal linking structure is a mess, your schema is missing or incorrect, none of the content work matters. AI crawlers use the same infrastructure as traditional search crawlers and they cannot cite what they cannot read.
What Does an AI SEO Service Look Like?
AI-Driven SEO vs. Hybrid Human/AI
The choice depends on budget, risk tolerance, and the specific goals of the campaign. Here is how the two models compare in practice:
| Hybrid Human/AI | Pure AI-Driven | |
|---|---|---|
| Who handles strategy | Human SEO strategist | Human at key checkpoints only |
| Who handles production | AI, with human QA and editorial pass | AI agents throughout |
| Best for | Most businesses; balances quality and cost | High-volume content needs with strong internal oversight |
| Content volume | Moderate to high | High |
| Cost | Mid-range | Lower |
| Risk | Low, with proper process | Medium: quality depends entirely on process quality |
| Suitable if you want to | Outsource the whole process | Maintain internal control with agency setup and training |
What neither model involves is asking ChatGPT to write articles and publishing whatever comes out. That is not AI SEO. That is the thing that gives AI SEO a bad name.
What Makes an AI SEO Agency Worth Paying For
Any business can access ChatGPT, Gemini, or Claude. What they cannot easily replicate is the accumulated knowledge of what makes content rank, the prompting frameworks built from years of SEO practice, the quality standards that distinguish useful content from generic content, and the technical implementation that ensures the content is properly structured, properly linked, and properly indexed. When you work with a professional AI SEO agency, you are paying for the expertise to use AI in a way that produces results, not just output.
Start with AI SEO Audit Checklist
If you want to assess where your site stands against the new requirements of AI-first search, here is a practical starting point. This is not a complete list, but it covers the areas that make the biggest difference.
Technical foundation:
- Is every important page crawlable and indexable? Check your robots.txt and meta robots tags.
- Are your Core Web Vitals in good shape? Slow pages lose visibility in both traditional and AI-powered search.
- Is your JavaScript rendering correctly? If your content is loaded client-side, check that Googlebot can see it.
- Do you have an XML sitemap that is up to date and submitted to Google Search Console?
Entity and structured data:
- Do you have Organisation schema on your homepage with accurate
name,url,logo, andsameAsproperties? - Are your key service pages marked up with Service or Product schema?
- Do you have an FAQ schema on pages that answer specific questions?
Content quality:
- Does each page target a specific search intent, not just a keyword?
- Does the content provide something that the top ten ranking pages do not? Original data, first-hand experience, a specific angle, a more complete answer?
- Are your headings structured logically, with H2s covering the main topic areas and H3s covering specific questions within them?
- Is the content written for people first? Would a reader find it genuinely useful, or does it read as though it was produced to satisfy a search engine?
Internal linking and topical authority:
- Does your site have a clear navigation and internal linking structure?
- Does every important page receive internal links from relevant pages elsewhere on the site?
- Are your internal link anchor texts descriptive and keyword-relevant?
GEO and AEO readiness:
- Are your key pages structured to answer specific questions directly, with the answer in the first paragraph rather than buried in the body?
- Do you have FAQ sections that mirror the conversational queries people use in AI tools?
- Is your brand mentioned consistently and accurately across external sources: Google Business Profile, LinkedIn, industry directories?
Conclusion
Every major shift, from the link-building era to the content marketing era to the technical SEO era, has caused people to worry about the future of SEO. This shift is bigger than the previous ones, because it is not just about how Google ranks pages. It is about what search is. The move from a list of links to a synthesised AI answer changes the relationship between content and visibility in a way that keyword optimisation alone cannot address.
The agencies that will do well are the ones that are honest with their clients about what has changed, that build the expertise to use AI tools properly rather than just cheaply, and that combine technical knowledge with genuine strategic thinking. The ones that will struggle are the ones that either pretend nothing has changed or hand everything to a language model without the oversight and expertise to make the output worth publishing.
The conversation about what SEO should look like (whether AI, hybrid, or traditional) always starts in the same place - being clear about what you are trying to achieve and what the right process is to get there.
About me:
I am the founder of Opace Digital Agency. I've been working in search engine optimisation since 2008 and managing projects for clients across the UK and internationally. Check out my profile on dev.to for more articles, or drop me a message 👇🏻
Top comments (0)