Maximizing SEO Impact with AI: Strategies & Tips for Success

Hey, it’s going well. Thank you.

Sure. No problem. Yeah. So I first became interested in digital marketing at secondary school when I was researching sort of potential career paths that didn’t require a university degree. By the time I finished secondary school, I was fed up with the traditional education system.

And then digital marketing took my interest as something I could pursue on my own in between my A level. So I started taking some of the Google ads certifications online and the Google analytics individual qualification. And then after I left school I started volunteering at discoverable so that I could learn on the job.

And then eventually that transformed into a full time position, which is good for me. In terms of the agency itself, we’re a small team we’re quite family run. And we have a lot of we focus on local clients. We have a variety of sort of small to medium sized businesses. And we really focus on fostering close relationships, transparent reporting.

And we just want to make sure that we’re doing what’s best for the client and making sure that, we’re driving the goals that matter to them most. So our main areas of expertise are SEO, PPC, which includes obviously paid social. We have an in house data analytics expert, and then also more recently just become part of my wheelhouse artificial intelligence, and then how that can be utilized within these areas as well.

Yeah. So my first exploration of artificial intelligence was really fueled by the public release of chat duty, I think sometime late in 2022 before that I didn’t really have any experience with AI.

I know it was around, but obviously that release did push it into the public eye a lot more. To begin with, it was really interesting to me just personally because I love learning about and getting involved with emerging technologies. But that did quickly transform into seeing what uses it might be able to have for my job in terms of SEO, in terms of in terms of digital marketing the industry obviously has seen the huge explosion since then in terms of open AI and chat GPT new LLMs being developed and released. And then obviously the side to that as well is developers creating their own sort of third party tools based off of these technologies as well. So there’s been a lot to keep up with and frankly, that’s almost a full time job on its own. I don’t really know anyone who’s probably kept up with it perfectly over the, over that time, but in terms of how I first started using it like many people, I imagine I first started with some pretty basic usages of the technology, that’s really necessary in my opinion to, to help learn a bit more how it works and what it’s about.

So for example, things like how can it, or what are the best ways to create high quality content, for example, using these tools at the beginning, I was just impressed that it even could just based on just based off of a prompt that it could write such good content. Then for example, probing the technology itself to find out how much it knew about SEO, fundamentally speaking and then also how well could it give personalized advice based on real world client examples.

When it came to, for example, technical issues, website development, and other sort of aspects of on page SEO. As more technology became available as well, I’ve more recently started experimenting with natural language to image generation models such as DALI Stable Diffusion, and others as well.

And then seeing if they are useful for digital marketing workflows, although I would definitely say that they have a bit more of a niche use case compared to the large language models themselves. Outside of maybe creating some images for blogs that sort of thing I haven’t personally found a great usage for them yet.

But again as the technology develops and we have new things like sora coming image to video which looks Ridiculous, by the way, I think that would definitely develop as well in terms of these cases.

Yeah, so it’s obviously, as it has developed a lot and they’ve been developing very rapidly over the last couple of years. It’s, the speed in terms of the growth in the response time, the features and overall capabilities has been pretty unprecedented. I could never have expected it to get so far so fast.

Which means the use cases for SEO and streamlining workflows are evolving constantly. And that’s something we, every time they release an update is something new to experiment with in terms of SEO and digital marketing. For me though, there are some definitely key areas where I use it most day to day and what’s had most significant impact for both our business and obviously our clients as well.

So the first thing, really one of the first things I use most day to day is it’s Google sheets capabilities. So one of the main things there is more complex formula creation. So creating formulas, which allow for large amounts of data manipulation which I mean, before we had tools like the ones that are available right now, I was spending more time than I care to admit trawling through forums, trying to find these formulas that did these specific things I wanted to do quickly.

And I’d spend more time trying to make it work than the time it probably would have spent doing these things manually. And now more often than not, ChatGPT can nail those formulas first time. Not always, 100%, but that goes without saying. And sometimes it needs a bit of nudging to get in the right direction.

But yeah, more often than not, that saves a significant amount of time, which is really important for, we at least spend, we’re in Google Sheets every single day and it’s a critical part of what we do. And the amount of really probably hundreds of hours saved Just on that one thing alone makes these tools absolutely invaluable.

Instead of having to hunt to where you’ve misplaced a comma in a, a 20 line formula, so that’s one thing, a hundred percent. And then the other facet to the Google Sheets the Google Sheets use is actually Google Apps Scripts. Which is a really powerful tool that I probably did not make enough use of before before sort of ChatGPT and the like.

Mostly because I’m not a developer and I don’t have any background, traditionally speaking, in developing in any language. But ChatGPT has opened the doors to doing more complex things there. An example being something I did fairly recently just a simple thing really, but something that isn’t as, accessible within just formulas, just things like equally distributing certain values throughout certain cells and then replacing dynamic things within each variation.

And this is just allows you to get a bit more a few more angles out of it that just aren’t inherently aren’t natively supported in Google sheets. So that’s a huge thing. For sure. Yeah, something I think is not as commonly talked about when it comes to using AI in SEO and digital marketing as a whole is the ability and it’s still, it is still coding based, which sort of aligns with the Google Apps Scripts thing but it’s creating custom tools that align to specific client needs and business goals.

It in a way it does fall outside of the traditional remit of digital marketing. In the, this is not really on page SEO and there’s not really anything like that, but at least in our agency, we do sometimes step outside those bounds. If it helps the business as a whole, we really try and get to the bottom of what these guys need the most, even if that doesn’t, even if that’s not something like Patrick page content or something just really, traditional.

So for example, just as some random example, creating a chatbot for a client where if they’d gone and outsource that to a specialized AI agency or a specialized chatbot agency that could cost, I don’t even know how much, but a lot of money but we can create something relatively simple, but actually it does the trick.

I actually also created one for our business internally and even does something you can select modes within it and then it roots off to an open AI chat GPT assistant, which I’ve created in the background so you can go between modes, which is pretty cool. And then just other things like I started experimenting with the content optimization tool, which sort of helps with keyword density and that sort of thing.

And these are things that I would just not be able to do from scratch to be honest. And although ChatGPT definitely has limitations when it comes to coding it’s not perfect a hundred percent. No, I think everyone who’s ever tried it would, knows that. But it’s good enough that if you have at least a basic working knowledge of things like JavaScripts and Python, you can probably get 85% of the way there and then use the knowledge that you have to fix the bugs and get over these little mounds.

Or, that’s when you go to the internet and get your head start.

Yeah, a hundred percent. And I think there are, obviously there are specific models that you can that are available and hugging face and such, which are more specifically niched down to coding. And I know there are individual tools as well, not that I’ve tried too many of them but that would definitely some, that’s definitely something that will improve.

And probably will become scarily good. It’s already pretty good as is. So that’s another big thing. So it’s just these little things where beforehand it would be like, yeah, I just can’t do that. There’s just no, no way I could even begin to do that. It’s now a case of actually, this could be possible.

Let me go and use AI to see if that’s something I can make a reality. There are a couple of other things which are really useful in terms of streamlining workflows and are very specific to AI. One being that ChatGPT offers fine tuning of its models directly within the OpenAI interface, and which is actually really powerful and something I don’t think is talked about enough.

Because it allows you to take all of the hard work that OpenAI have done in terms of training their models to get where they are today, and then applying your own very specific use cases on them. Something common that people could do a lot, I’m sure, in this industry is tone of voice adherence. So you can create client specific models that, when asked to create content, But for that model, it generates it in that client’s turn of voice because you’ve trained it on their content.

There are some nuances and some quirks that fine tuning has, which you have to experiment with to get good results. But once you’ve nailed it, it’s something that can really help. And the same goes for structure of content as well. So you can train it on specific structures or specific content lengths.

And that is something that I think a lot of people could make good use of.

Yeah, so there’s actually a lot of misconceptions that I still hear about AI. And although it’s been out for, although the very concept of AI has been around for many years it’s still a relatively emerging and new technology. So it makes sense that these, there are these misconceptions.

On the one hand, an opinion I hear quite often is that AI has absolutely no place in SEO, which is one, one side of the coin. And that it has no practical applications. Although I do actually personally think it’s quite a narrow minded viewpoint, I do completely understand why these viewpoints exist.

With the current capabilities of these technologies, albeit amazing, there are some fundamental aspects of SEO that can’t be undertaken by these models. One example I can think of is some parts of technical SEO auditing. It can’t perform these real time these real time tests on on pages. Large scale consistent content auditing and actioning. Just because, although it’s good at handling large amounts of data, it’s hard to create reproducible results that are to the standard of, obviously, an expert in SEO who’s been doing it for years. So there’s things that you can’t do a hundred percent. In my mind though, the trick is understanding where the opportunities lie within the boundaries of large language models as they currently stand. That being said, there are some people with this misconception that just suffer from bad prompting for their needs, which is a bit different, and that’s something that can be fixed. One of my core beliefs when it comes to using tools like ChatGPT or Gemini is that the quality of the outputs you receive are directly proportional to the quality of the input you give it. So earlier this year, I actually helped contribute to an article on prompt engineering for the economist.

And it’s something I really was really passionate to do because prompt engineering is one of the most underrated aspects of how to get the best out of these models. A hundred percent, because I still all the time see people using very basic prompts, but they either one don’t include all of the necessary information needed.

Two, it includes no examples where it would be beneficial to do which we would call few shot prompting. Or three, which is a less commonly talked about, but in my mind is still very important that the prompt they give it actually biases the response. So if you nudge ChatGPT, for example, towards a certain response that you want it to give, it won’t hesitate to give you the response you want, even if it maybe isn’t the case or just isn’t true at all.

So it can lead you down the garden path if you’re not careful. There’s a lot to consider when it comes to prompting. And I think in a lot of these cases, if people improve their prompts, I think they’d realize that there was actually a lot more that these models can provide in terms of value.

I think really, for this misconception, the main takeaway is that AI absolutely can have a place in SEO. But it’s important to understand that the key to unlocking the full potential is you’ve got to use it as a tool, not as, not as a standalone thing. It has to be alongside your experience and expertise in, in this.

It can’t be just, you can’t just give it some SEO tasks and it goes and does them. There is something, there is another misconception I hear, but it’s it’s similar, but it’s a slightly different angle. Some people think that AI has every place in SEO, but so much so that just you don’t need humans whatsoever. So some people say you shouldn’t use it at all. And then some people say, actually, it doesn’t need to do anything. We don’t need to do anything anymore. AI has been taken care of. And that is, yeah, exactly. So that is something that I see a lot, I obviously I’m a bit, I do have a horse in this race.

I don’t want to lose my job to the robots anytime soon. But I do strongly believe that we’re nowhere near the reality of the, these tools being able to do SEO, like just, I do know there are some tools out there that claim to do that. And undoubtedly there are some things they can do a hundred percent. There is a use there. But I’ve tested, a lot of these stuff towards myself. I’ve tried to automate some elements of SEO using AI and large language models. And the realistic outcome is that I, there’s this really hard to produce consistent enough resultsthat you can be confident relying on them when it comes to implementing a content strategy. The concern that I have is that these opinions are often found in startup businesses, by small business owners, or in sort of family run side hustles, where they may not have the budget to, to ask an SEO expert or to go to an agency.

And they’re really having to rely on, on, on these tools because there’s no one else to ask. And in those cases, it could be that this advice is doing more harm than good, because even in the cases where a chat GPT may give a disclaimer saying, I can’t do everything, go and speak to an expert. It’s my experience that most people don’t do that. They’d take that disclaimer and don’t follow it. So I think, It’s, it just comes down to a balance. It can be used, but it’s important to understand what it can’t do as well. And then finding the balance between those two things.

That’s a tough one. There’s so many. But I think really my absolute favorite instead of a favorite client story, I have a favorite thing that I do enjoy doing for clients. And that is more in that creative thinking aspect. It’s not just conforming to these. To these core SEO strategies and these core principles that we follow. And obviously they’re very important, sometimes they can get a bit samey, a bit tedious, you know how it is. So I enjoy coming up with these creative solutions to problems. Maybe it doesn’t, maybe it’s not traditional SEO. Maybe it’s a case of a client needs a smart solution to product management or product categorization. And I really enjoy using tools like chat GPT to come up with a smarter solution to help them save time. In some cases, a lot of time and money doing something that actually can be achieved using AI and using our experience.

In just a smarter way than paying for these expensive often pre made solutions by third party companies.

You’re very welcome. It was a pleasure to be here.