AI can help your business optimise its SEO efforts by providing suggestions for improving content and technical issues, optimising on-page copy, and other factors that impact search engine rankings.
Nonetheless, these tools are just a starting point. Here we look at the pros and cons of using AI for SEO and whether relying solely on those is enough…
What does AI mean for businesses?
If you want your business to be successful online, you’ll need to be ranking in the top 10 search results on Google. According to Semrush, the #1 result in Google gets 27.6% of all clicks, and moving up one spot in the search results will increase CTR (click-through rate) by 2.8%. This means more clicks and ultimately conversions.
If you want to capture this valuable search traffic, you have to optimise your website using various SEO techniques.
AI tools can help you to do just that, by identifying areas for improvement and providing you with suggestions to improve your SEO. According to PWC, business leaders believe AI is going to be fundamental in the future. In fact, 72% termed it a “business advantage.”
But, specifically, how can you use AI for SEO?
Keyword Research
Do you need to find out what people are searching for online and come up with ideas for a new blog? Do you want to understand your competitive landscape and how difficult or easy it will be to rank in the top spot on Google?
Keyword research is all about finding the right terms that your target audience is searching for to find information, products or services online. According to Ahrefs,
94.74% of keywords get 10 monthly searches or fewer
0.0008% of keywords get more than 100,000 monthly searches
Therefore, the key is to finding the right keywords that are easier to rank for and convert better, which are usually long-tail, question or phrase-based terms. Here we explain how AI tools can help you find the right ones…
How to use the AI tools for keyword research
Imagine you’re a start-up selling a unique blend of flavoured tofu. You’re writing copy for your e-commerce products to ensure you can compete with the big guys. You’re also looking for content ideas to create blogs/recipes and link them through to your products.
Unlike Google Keyword Planner, Ahrefs or Semrush, AI tools such as ChatGPT do not have access to search volume or other metrics, such as website performance data.
Nonetheless, with the right prompts and guidance, it is a good starting point to build a list of terms that might be relevant to your business. Here is how…
1. Create an initial list of seed keywords analysing your website and/or products/services you offer.
Seed keywords are words that act as a starting point in keyword research to unlock more keywords. For example, if you sell tofu, then you can use seed keywords like tofu, soya beans, marinated tofu, etc.
To create a list of those seed keywords, you can have a brainstorming or ideation session which would allow you to come up with relevant terms that describe your business, product and/or service.
You can also use ChatGPT as a starting point, using a prompt like the one below:
2. Research keywords
Now that you have your seed keywords, you need to identify some keywords that fall into those main categories. You could use tools such as Chat GPT to come up with a list of keyword ideas that could be useful for content writing, product descriptions and more. Here is a cheat sheet from Ahrefs listing the types of keywords you may need to explore.
[ Source: https://ahrefs.com/blog/types-of-keywords/]
This is also where AI tools such as Chat GPT can be seen as an advantage. All your competitors are using more or less the same keyword tools and using Chat GPT could help you come up with more creative ideas.
Nonetheless, it’s important to remember that with regard to the output, the prompts are equally important. Here is an example:
This list can simply function as a starting point to build upon. Of course, you could dive deeper into the topic and play around with the prompts.
3. Look into topic clusters to ensure your content is organised and relevant to the subject matter.
Let’s use the topic of “marinated tofu” as an example. You would need a pillar page on marinated tofu that captures all the key aspects in a single page. The additional pages that would form a cluster could relate to stir fry recipes – here is an example of keyword ideas using free tools such as Also Asked or Answer the Public but please note those are limited to a few searches per day. Here is an example from Also Asked.
ChatGPT could also help with creative topic ideas and categories that could apply to your business.
4. Understand how search intent impacts keyword research
Search intent is the reason behind a search query, i.e. why a user types a word or phrase into a search bar. We have discussed search intent in detail here.
Let’s say, for example, you’re researching the keyword “how to marinate tofu” for an article you want to create.
Does the user searching for this query want to learn how to marinate tofu? Or do they want to buy marinated tofu?
If your content strategy is only targeting people interested in the latter, you’ll need to expand your content and include informational pieces that would engage people at the beginning of their buying journey.
From the search results, we can see that for this term, there are blogs and recipes that rank in the top positions.
As a result, it is important that your website solves the problems users are researching online, rather than simply featuring the keywords that are searched for.
5. Spot-check the data
Finding specific keywords to target is just the first part of the journey. When using tools such as Chat GPT, always spot-check to weed out irrelevant queries that don’t apply to your business.
6. Research the search volumes, keyword difficulty, trends and competitor performance
As a part of keyword research, you also need to look at supporting data such as search volumes and keyword difficulty – this will give you an idea of the popularity of various keywords and how easy or difficult it will be to rank for them.
In our experience, this is where Chat GPT falls flat. The tool is not coming up with a list of ready-to-use keywords, but keyword ideas that need to be cross-referenced with your business offering and checked against other data such as search volume, keyword difficulty, trends, competitors etc.
ChatGPT keyword research limitations include:
- ChatGPT does not have direct access to keyword data and can’t perform keyword research that would look at different aspects mentioned above from keyword ideas to search volumes and competitor data
- The platform is known to provide information that is out of date
- ChatGPT tends to break a lot which hampers productivity
- Chat GPT has only been trained on data up to 2021 which means it may miss certain terms/words that were trending before this date but might be still relevant
- ChatGPT requires detailed prompts and will be driven only by the person who asks the question. This means that if you’re not aware of certain aspects or best practices around SEO, you may subsequently receive information that is not relevant or in line with the latest guidance.
How to overcome ChatGPT limitations:
- Check any keyword suggestions against search and competition data using tools such as Semrush or Ahrefs – these are paid for tools which may also require training in terms of how to use it and understanding the data as well as its impact on performance.
- Depending on your budget or availability, your preference may be to partner with a digital marketing agency that not only can provide strategic guidance but also actionable insights based on in-depth research.
Meta Data Writing & On Page Content Optimisation
Are you looking to build a new startup website or make your current site more effective?
Although AI can help you create a variety of content, including blog and on-page content, as well as meta data such as H1 headings, title tags and meta descriptions, solely relying on those tools may present new challenges.
Google has recently reiterated guidance on AI-generated content and stressed the importance of original, high-quality, people-first content demonstrating qualities of expertise, experience, authoritativeness, and trustworthiness. Therefore, although it’s useful for creating meta descriptions or at least speeding up the process, it’s important to know its limitations where the output is often repetitive, out of date or not factual.
Example:
I have asked AI to write a title tag, a meta title and a description for a product that is sold online – I used tofu in this example.
Although it’s ok to use as an inspiration, the written text is clearly longer than the recommended character limits. Consequently, when your meta tag is too long, Google can cut it off or replace it with its own description.
This is an example where, as a business, you can’t blindly rely on AI recommendations and check everything before applying any changes to your site. Using the above suggestions will mean that your meta title is over the recommended limit.
How to use AI for On Page Content Optimisation
- Start with keyword research to ensure the relevant keywords can be included in your meta data and throughout the copy.
- Make sure your content aligns with search intent (I discussed it when talking about keyword research)
- Make sure your content is easy to read and organised well
Your content needs to have good headers (HTML element <h1>, <h2>, <h3>, and so on) which are positioned throughout your page to organise content and help search engines understand what your content is about.
To create those, ChatGPT would be useful too, however looking at the below, it’s important to ensure the instructions are very detailed.
It’s good practice to incorporate important keywords in your headers naturally, especially in your <h1> and <h2> headers. Also, the headers need to be written in line with search intent to ensure Google values your content and ranks it higher.
Providing another prompt produces better results but these are quite repetitive and boring. This is where the limitations of ChatGPT come into place and the importance of using other tools or partnering with an agency comes to light.
4. Make sure your content has relevant and persuasive calls to action which
- Incorporate emotional appeal referring to the prospect’s pain or goal that triggers action.
- It’s all about timing aligning with the buyer’s journey and adding urgency when needed.
When asked, ChatGPT does provide some good options, however, those need to be reviewed and considered in line with the type of content or whether they need to be applied to the buttons on the site, in which case will be too long.
5. Ask AI to write a title tag, meta title and meta description for the newly created piece of content or an existing piece which you want to optimise. Specify the character limit.
Meta title/title tag – 50-60 characters
Meta description – up to 160 characters
Did we mention that the tool is not reliable and breaks often?! Let’s try again.
Although it rectified the issue, you may want to be more creative and use some unique selling points in your title tag such as unique flavours or speedy delivery options. For this, you could edit the prompt or include your own creative flair.
Don’t forget to put all the meta data in the spreadsheet which will help you count the characters using the =LEN(cell) formula.
6. Always spot-check and read through to ensure the content is relevant to the page you’re describing using the meta data.
The limitations of ChatGPT for on-page optimisation:
- ChatGPT’s content sounds robotic at times, lacking the understanding of idioms or phrases with a figurative meaning.
- ChatGPT can’t provide deeper insights into the true meaning or reasoning behind users’ search queries. It can provide facts, although some are out of date and questionable, but it won’t dive deeper into human emotion and the true drivers behind some behaviours.
- ChatGPT is still learning and its answers will be changing continuously.
- ChatGPT is not human and won’t provide you with the holistic recommendations and actionable insights into SEO optimisation that an agency or a marketing consultant can do.
Technical SEO
Do you want to make sure that your technical SEO is up to speed? Do you want to ensure that your website is easy to navigate and is free of any technical issues that prevent it from being ranked by search engines?
In addition to helping with on-page content, AI SEO tools can help with technical SEO efforts, from detecting missing meta data to managing canonical tags and SEO forecasting.
Paid for tools such as Market Brew, use AI to analyse the future of your rankings and help determine your strengths, shortcomings, and trends. This is based on a number of audits and processes that the software monitors in real-life, from keyword rankings to link value and topic authority.
[Image created by MarketBrew.ai]
Similarly, Alli AI can conduct real-time website audits, flagging issues that need your attention.
[Image created by alliai.com]
Can I use Chat GPT for technical SEO?
Depending on your level of knowledge around technical SEO, Chat GPT can help to identify the areas that come under the term ‘technical seo’.
However, this is where we come across the limitations of the tool.
If you’re not well-versed in the technical aspects of your website, it’s good to speak to a professional marketing agency that can help you identify any issues mentioned above and provide fixes. An agency will also be able to provide strategic recommendations and best practices to ensure that moving forward those issues are avoided, for example with regard to URL structure or internal linking.
With the failed Core Web Vitals, for example, it’s important to optimise the site for web speed due to the fact that this is now a ranking factor. If your content is of the same quality as competitors’, Google will rank their site higher if their mobile and desktop speed is better optimised.
The Limitations of ChatGPT for Technical SEO
- Chat GPT has only been trained on data up to 2021 and its responses will be generated based on this data, it is not an SEO expert and the output may need to be reviewed and edited by a human expert.
- ChatGPT cannot directly access external websites, therefore it can’t identify any technical issues with the website.
User Experience
Do you want to increase conversion rates across the site? Do you want users to engage better with your site?
In line with recent Google updates, Page Experience has an impact on SEO. In fact, 38% of people will stop engaging with a website if the content and layout are not user-friendly.
Take Netflix as an example. Netflix has a constantly changing AI algorithm that helps improve content recommendations.
In theory, the same goes for your website.
Personalising user experience and showing them exactly what they are looking for will improve conversion rates and engagement metrics. This will impact rankings.
Example:
Tools such as Evolve AI, allow for creating personalising user journeys by continuously adapting to live user behaviour. Market Brew, which we mentioned above, can mimic search engines and provide recommendations on what you can do to improve your website’s UX.
If you’re a startup you can use tools such as Teleport to design your website and play with UX variations. How about the below? Great results in under 5 mins.
Nonetheless, as we mentioned previously, human expertise is necessary to ensure there is someone to manage the project, whether it’s a web design or optimisation of the current site. Although data can help you make better decisions, it all depends on the available budgets as well as overall strategy.
Find out more about website design for a start-up here.
Can I use ChatGPT for UX?
Use this tool to help you identify the most important aspects of UX.
The good news is that some of those recommendations can be applied internally, for example, the use of calls to action. If budgets are tight, you can play with different colours and content to see which ones are performing the best. It’s also important to ensure you can track those using tools such as Google Analytics. This will help you to make a decision with regard to the calls to action that bring the best results.
Sure you can use those across your site, however, it’s necessary to bear search intent in mind as well as where in the buyer journey your users are. Whereas the above are great for commercial intent for users that are ready to buy, for those who are researching different vendors, ‘subscribe’ would be a better option.
What are the limitations of ChatGPT for UX?
- ChatGPT itself is not capable of collecting data to analyse the performance of your UX efforts, for example, to decide whether a certain call to action works better than the other.
- ChatGPT lacks the understanding of human behaviour and emotion and depending on the subject matter may not seem empathetic or recommend changes in line with cultural and social factors that are so important to us humans. From the choice of font to the visual elements, the tool will struggle with the design suggestions that effectively engage the target audience.
- ChatGPT is not human and wouldn’t be able to cooperate and take into consideration stakeholders, audience, developers and others involved in the project.
Although ChatGPT is an impressive AI language model, for UX design it definitely falls short. Without a deeper understanding of human emotion, and the ability to collaborate effectively and process data, the tool fails to provide the critical thinking and creative decision-making that underpins all kinds of user research.
When it comes to understanding the depth of a user’s needs, feelings, and desires when experiencing your products or services, the human factor cannot be replaced. This coupled with the use of the right tools such as Google Analytics or Hotjar to help you understand how users behave on your site is the perfect combination when working on any UX project.
Proceed with caution
The capabilities of AI tools are tofu-tally awesome (we actually used AI to come up with this). Nonetheless, there are plenty of practical issues, and while those tools are incredibly useful it is important to remain cautious and use them in conjunction with other tools, as well as human input.
For example, apart from those above, we have observed instances when the tool provided outdated information, as well as repetitive content. Moreover, individual company values aren’t reflected, as well as the tone of voice.
If you have any questions about AI tools, or would like to know more about any of our other services, please get in touch with our team of experts.