There’s a lot of talk in the industry about AI and how it helps SEO, how it’s being used in various scenarios, and even the changes we see in SERPs with AI Overviews. 

AI is everywhere at the moment, and we even see most of our everyday tools incorporating artificial intelligence into their offerings, such as AI Writing tools, Automated Keyword research, and alt-text generation, etc. 

But here’s where I stand on it all…

AI should be a tool, not a strategy.

If you rely entirely on AI for SEO, you're missing the finer details that only human expertise can detect.

Nikki Halliwell

Where AI Helps SEO (But Falls Short)

Let’s look at some of the most common areas where I see AI cropping up and even being recommended by others in the industry. 

Keyword Research

AI suggests keywords based on patterns, but it doesn’t know your industry’s nuances or real search intent. You will often get thousands of keywords with no relevance to your business or industry, and the data is often entirely wrong. AI cannot access search volume and similar data and is confidently wrong in most of its outputs. 

Manual research ensures we get high-quality, relevant keywords that are validated by business insights and can be mapped to the necessary intent and business functions. We also have several tools at our disposal to validate the data. 

Content Optimisation and Generation

AI tools analyse NLP and semantic keywords and can speed up mass page title and meta description optimisations. However, they can’t replace human storytelling, brand voice, or strategic messaging. 

Chatbots and LLMs really struggle to add or recommend relevant and useful internal or external links. They also have a known problem with correctly citing sources.

Technical SEO and Site Audits

As you might expect, this is where I have the biggest issue with AI in SEO. 

AI audits may be able to point out specific errors, but they won’t see conflicts in internal linking or UX problems that impact rankings.

An SEO audit is not an export of issues from an SEO tool with no additional context. A tech audit is not a list of URLs attached to an issue or even a checklist with a top-level summary. 

It should be a holistic view of how all aspects of the website are performing and how they may be affecting the ability to perform and compete in organic search results. 

Don’t get me wrong, AI is useful, but it cannot convey how website issues can affect business KPIs and revenue.

No technical SEO audit is complete without context and detailed recommendations on how to address the issues you came across. It is also not a one-size-fits-all approach, so the recommendations need to be tailored to each website and take into consideration unique limitations and blockers that the business may encounter. It may even need to include contingencies and plan B, etc.

With the best will in the world, this is something that LLMs cannot do and should not be used in this way.

Algorithm Updates & SEO Strategy

AI can pretend to react to ranking shifts but can’t make educated recommendations based on previous actions and the upcoming pipeline. As SEOs working closely with our clients, we know how to adapt based on experience, industry insights, and testing.

LLMs are trained on old data, so by the time they suggest what someone should do next in reaction to a drop in performance, the landscape has likely already changed, or they are recommending outdated practices. 

AI vs. SEO: Why Human SEO Still Wins

Does AI Have a Place in SEO?

Yes, within reason. 

I do use LLMs in my daily work, but they don’t do the work for me. They take away some of the repeated tasks, leaving more time for me to do the real-SEO work that my clients pay me to do. 

AI is useful, but SEO isn’t just data—it’s strategy, creativity, insight and expertise.

The best approach is to use AI for efficiency but always rely on manual SEO for accuracy, context, and long-term success.

Be wary of those trying to sell fully-automated solutions or audits, and always speak directly to the person behind the computer screen so you can be sure of who you’re working with. 

When optimising page speed, work on the page templates rather than individual pages. By focusing on the core templates, you can implement improvements at scale, ensuring a consistent boost in page speed across various pages.

This approach enables easier collaboration and support from stakeholders who recognise the overarching impact on user satisfaction and business objectives.

The Power of Page Templates

Targeting page templates for optimisation means addressing the root structure that influences multiple pages.

Instead of changing one page at a time, we can implement fixes across a range of URLs that are built using the same design templates.

This centralised approach to optimising site speed allows for efficiency and consistency in implementing your improvements.

Scalability of Improvements

Ensuring that improvements are applied uniformly across pages sharing a common template means that we create a consistent and streamlined user experience. 

This is significant as it allows us to reinforce more positive interactions that our audience has across the entire website.

Stakeholder Buy-In

Collaboration becomes more seamless when stakeholders recognise that site speed fixes made to page templates benefit the entire website.

When stakeholders see the positive impact across a range of URLs that use the same template, they are much more likely to buy into further optimisation work and support you even more with your next set of recommendations. 

Deciding which Templates to Focus on

When you’re faced with a large website, you need to know which templates you should begin with. 

Begin by identifying the core page templates that are the basis for your website’s content. This may include a homepage, product pages, category pages, etc.

Next, we can conduct a comprehensive site speed audit of each identified template, analysing elements contributing to load time, such as images, scripts, and server responses.

Utilise performance testing tools like Google PageSpeed Insights, Lighthouse, or GTmetrix to identify specific areas for improvement within your page templates. From there, we should be able to see which of your templates is the slowest loading. I would typically focus on improving that template first.

Monitoring Pages and Improving Performance

After sharing your template-specific recommendations with stakeholders, I suggest implementing a system for continuously monitoring page speed metrics, especially for the optimised templates.

Regularly assess the impact of your changes and refine your chosen strategies based on ongoing performance and changes to Google algorithms.

This monitoring approach ensures efficiency at scale and produces a more collaborative atmosphere. This way, stakeholders tend to quickly rally around the shared goal of delivering a seamless and fast user experience on their website.