AI has been at the top of everyone’s mind for about a year. The massive success of ChatGPT has put the topic on the agenda of multinational companies and startups alike. Many brands, however, struggle to effectively work AI into their workflows.
Initially, many believed that AI would be a great shortcut to create content at scale. Blog posts, product page listing copy, you name it. Many people considered AI solutions and ChatGPT in particular as a shortcut to ramping up content output and to reap the benefits of improved traffic from Google search results.
The Rise and Fall of AI-generated Content
Most brands that used this tactic have since regretted it. While many SEO case studies show that AI content can in fact rank fast and can increase traffic from Google search results in the short-term, the results often don’t last.
The screenshot below is from a case study that Martin Jeffrey published on LinkedIn. The case study is based on a test website that is populated with AI-generated content. As you can see, organic traffic starts ramping up initially, but then sharply drops. At Impala Services we have seen this pattern repeatedly and we are consequently cautious when it comes to AI-generated content.
Google’s Point of View on AI-generated Content
Over the last few years, Google has given search engine marketers consistent guidance on the type of content that will rank on the Google search results page: Our focus on the quality of the content, rather than how it is produced, is a useful guide that has helped us deliver reliable, high-quality results to users for years. More specifically, Google is promoting its E-E-A-T framework to assess content quality: Expertise, Experience, Authoritativeness, and Trustworthiness. This is where the issue with AI-generated content starts: while it can convey expertise, it is not based on any form of more profound research, or analysis. It is not written by an authoritative author. It often does not include references to other authoritative sources. In conclusion, Google stops short of banning all AI generated content, by stating: “This said, it's important to recognize that not all use of automation, including AI generation, is spam.”
The AI Content Sweet Spot: Short and Generic Copy
With these cautions in mind, there are many cases where AI can be used to at least draft highly tactical and short copy.
Using AI for Page Title and Meta Description Copywriting
A page title is a piece of text in the HTML code of a web page. It is often not displayed on the page itself but is identified and used by Google crawlers to generate the Google search page listing of a website. It should be succinctly written, describe the purpose of a page, and be between 50 to 60 characters long. In other words, it is pretty generic copy and it is often not possible to manually write the page titles for every single page, especially for websites with thousands of pages. In the past, many website page titles were auto generated using simple rules, such as: the name of a brand plus the headline of a page. While this worked, in many cases the page title copy was not great and lacked specifics such as product details. The same applies to meta descriptions. Just like page titles, meta descriptions are not visible on a page, but extracted from its HTML code. They are the second component that goes into a Google search results page listing.
This is a great example for how to use AI in copy writing. AI can typically generate page titles and meta descriptions at scale that are much less generic and contain more relevant content than copy that is written using a simple rule-based approach.
Writing a Post Outline
When you get started writing content, such as a blog post, you need to make sure that you include all relevant sections that Google (and users) may expect to be included in the blog post. ChatGPT and other AI-based content creation tools can be very effective in making sure that you do not miss any important sections. So rather than using ChatGPT to write your entire blog posts, you ask ChatGPT to write you an outline and to give you section headlines. You can also provide competing pages that currently rank well for your target keywords and include them in your ChatGPT prompt.
Tweaking Existing Copy with AI
Rather than using AI to write new copy from scratch, we often use AI to enhance and improve human-written copy. AI can be effective at sharpening a dull and lengthy opening paragraph or giving you more ideas for section headlines. You can combine this with a prompt that compares content that currently ranks in a high position for your target keyword. This way, AI acts as a supercharged version of Grammarly. Consider that the purpose of the content on a page is to answer the question a user has (the search query). The more succinct and precise your content answers that question, the more likely your page is to be in a top position on the Google search results page.
Recap: Use AI Wisely and to Enhance Generic and Existing Copy
At Impala Services we are always on the forefront when it comes to adapting new technologies. That’s why leading brands like Amazon Basics, Lidl, and Kingfisher choose to work with us. We are also cautious and prefer to fully understand all implications of new technology before providing it at scale. For AI-generated content, we have found a way to provide our clients with the best of both worlds: fast and cost-effective copy generation via the use of ChatGPT with human oversight to ensure that the results last.
Other related posts you may also find interesting: The Importance of 3D Modelling in Advertising, How to Leverage Data Visualisation to Drive Sales by Improving Product Infographics
.png)
