5 Guiding Principles for Using AI in SEO
Pretty much every one of us in digital marketing and SEO uses AI for our work now. We know it can speed things up, get rid of manual and repetitive processes, and really amplify our work as SEOs. We can deliver more, and especially give clients more strategic thinking, instead of spending hours and hours on manual work like pulling data or writing meta descriptions.
But despite all the positives, it’s very easy to become reliant on AI, or to use platforms like Claude or ChatGPT the wrong way.
In this article, I’m going to go through some guiding principles for using AI the right way, so it amplifies our work.
1. We direct AI, AI doesn’t direct us
The first principle is that we should be directing AI. AI should not be directing us.
To get the most out of AI and really amplify our work (whether that’s as an SEO consultant or GEO consultant), we need to be the subject experts. Then we can direct it, get the most out of it and steer it in the right direction. If we’re not, it’s very easy to assume whatever a model tells us is correct and just go along with it.
In practice, that means we decide what AI works on and how it goes about it. It’s easy to slip into letting AI set the agenda, where we end up working on whatever it suggests next. That’s rarely the best use of it, or of us.
2. Batch your AI work and avoid context switching
The second principle is about multitasking and context switching.
Outside of AI, one of the basic rules of being productive is to minimise context switching and batch your tasks. With task batching, you might clear all your emails in one go, for example, rather than spending five minutes on emails, then five minutes on something else, and so on.
Because AI lets us do things very quickly, and do several at once, we need to carry that principle over.
Last week, for example, I had Claude working on several very different tasks at the same time:
- a fairly complex data pull
- proofreading a client document
- diagnosing some Google Tag Manager issues
By the end of the week, I realised that approach didn’t work. In some ways I was using AI effectively, because it was helping me get work done faster. But I wasn’t just switching between clients; I was switching between very different types of task. All that context switching was cognitively draining, so I wasn’t properly focused. I wasn’t using AI as productively as I could, and ironically, I wasn’t doing my job as well as I could either.
AI models let us work on different tasks, projects and clients at the same time, but we still need to stay focused as humans. That might mean sticking to one client at a time, even if the tasks vary, or working across different clients on the same kind of task.
3. Apply critical thinking
The third principle is to apply critical thinking. That means not assuming AI is always correct.
We also need to remember that AI doesn’t think, and it can’t always understand context. It can be very good at pulling data, sometimes more accurately than we would doing it manually. But when it’s giving opinions or influencing decisions, we need to give it our full attention.
The more expert you are in your area, and the more you work with AI and a particular model, the better you get at knowing, almost instinctively, what it’s good at and what it isn’t.
We need to be very mindful of this, especially when we’re giving a client something that could affect their website’s revenue or conversions. In those cases it’s vital that we bring a level of critical thinking, and some cynicism, to whatever AI gives us.
4. Don’t trust AI with your SEO strategy
I’m a great believer that AI is a powerful tool. It amplifies our work and means we can give clients better value. But I don’t believe AI is capable of delivering a full, robust SEO strategy.
A proper SEO strategy involves lots of different pieces, including:
- content
- digital PR
- competitor analysis
- technical SEO
- accessibility
- working with brand
- gap analysis
- analytics
AI can certainly help with all of that, but I’d never be confident in a strategy developed entirely by AI.
A well-rounded SEO strategy also takes into account the client’s circumstances and the resources they have. For example, some clients don’t have the resources to create much content, and some don’t have big budgets for digital PR. Any SEO strategy worth its salt is bespoke to the client: their industry, their needs, and what’s actually possible for them.
5. Don’t forget the soft skills
The fifth principle is about soft skills.
Whatever AI can or can’t do when it comes to collating data or helping with strategy, it can’t communicate effectively with humans. An SEO strategy could be really good, but if you can’t explain it to a client, be persuasive, and help them be persuasive with their own stakeholders, it never gets actioned.
Going back to the idea of us being the experts: yes, we need to be that. But we also need to be able to explain things clearly, sometimes in layman’s terms, and to be persuasive about our strategies. For example, that might mean talking to clients about how they can improve conversions and revenue, not just traffic, rather than only talking about keyword volume. It also means being able to explain why some (often boring) technical fixes matter.
It’s this human side, like presenting something complex to stakeholders, that makes the difference.

Andy Killworth
I’m a highly experienced digital marketing strategist offering a holistic approach including SEO, GEO, analytics and conversion rate optimisation.
Having worked across many industries including highly competitive ones like iGaming and finance, I have extensive knowledge of sites of all sizes and situations.
I also write openly about my own mental health, including living with bipolar and ADHD, in the hope that it helps someone else feel less alone.
I love running ultra marathons, spending time with my three kids, and making bad dad jokes.

