Top Tips for Generative Engine Optimisation 

By Ushna Prema

Generative AI has arrived, and it looks like it’s here to stay. While artificial intelligence and machine learning have been used for years within the life science and biotech industries for ‘big data’ analysis, drug modelling and making clinical predictions, the launch of ChatGPT in November 2022 saw a huge influx of accessible and free generative AI chatbots such as Google Gemini and Bing Copilot.  

Gen-AI chatbots, unlike conventional search engines, can find and condense information almost instantly, avoiding the need for users to search on multiple web pages. Not surprisingly, these effective and time-saving AI tools have garnered worldwide popularity quickly, with ChatGPT now serving over 120 million users per month (Duarte, 2024). 

Moreover, we are seeing a significant shift in the type of information users are requesting, with the ability of gen-AI chatbots to answer more detailed and executional search queries such as performing tasks. For example, we asked ChatGPT to provide us with a table of reagents required for cell culture: 

While the summarised answer was quite general, it did provide the information in the correct format, as well as highlighting the purpose of each product.  

Despite this, while gen-AI may be able to answer questions like “Where is the best Italian restaurant near me?”, more complex queries requiring summaries of scientific literature and research have been shown to result in ‘hallucinations’, which is where the AI-chatbot will give answers containing false, misleading or nonsensical information (Alkaissi and McFarlane, 2023). This is often perpetuated by the lack of citations or sources linked, making it difficult for users to make informed decisions on whether the information is true. 

What are generative search engines? 

As generative AI evolves further, chatbots are now being integrated into search engines such as Perplexity AI and ChatGPT Search (released by OpenAI). These generative search engines retrieve information from multiple internet sources and then use large language models to generate a summarised answer to your question. They offer several advantages that provide a more personalised user experience: 

• Conversational: Use of more natural ‘human language’ in their responses 

• Detail: If a response is not satisfactory, users can refine their search by providing additional information 

• Context: Retaining information from prior questions, together with data from cookies, browsing history and user search patterns to generate user profiles (Xiong et al., 2024) 

• Real-time data: Access to live data enables up-to-date sources of information to be used (unlike older generations of gen-AI chatbots which rely on dated sources). 

What factors impact generative AI ranking your website? 

The importance of website rankings cannot be overestimated, with the first five results on a SERP receiving 67% of all clicks (Radić, 2022). As we embark on a new era that utilises generative-AI in our everyday lives, how can you ensure you stay on AI’s good side? The answer is to optimise your website content to ensure your brand is found by both gen-AI chatbots and search engines – a new term coined generative engine optimisation (GEO).  

A study of 100 search queries in ChatGPT identified the following six factors that impact whether ChatGPT would be likely to recommend your product or service (Patel, 2024): 

Frequency of brand name: The more times the brand or product was mentioned on the website, the higher the chance it was recommended by ChatGPT 

Keywords: As with conventional SEO, the keywords used for a search query in ChatGPT matched those found in the source code of webpages 

Age of services: Products and companies with more established websites were recommended more frequently by ChatGPT 

Reviews and recommendations: Companies with reviews or recommendations on other websites tended to be favoured in ChatGPT answers. For example, our question “What are the top 10 DNA sequencing services in the UK?” produced the following results: 

In this instance, ChatGPT pulled information from an article entitled “Top 73 Genome Analysis startups” (https://www.medicalstartups.org/top/genome/) to provide an answer. 

Authority: Companies with greater domain authorities and larger followings on social media platforms were more likely to be recommended. 

Recent work looking at GEO highlighted the importance of domain-specific optimisation, whereby certain factors impact gen-AI engines’ answers based on the related industry. For search queries related to science, technology & healthcare for instance, generative-AI has been shown to prioritise authoritative language style, fluency of content, use of technical terms and inclusion of statistics in website content (Aggarwal et al., 2024). This style of writing would cater towards a scientific audience who typically like evidence to support new findings and results. 

Tips for optimising your content for GEO 

• Structure your website content so that it answers the questions users may ask, by incorporating long-tail keywords and organising text in a way that is easy to read. 

• Understand user intent behind search queries by providing comprehensive and in-depth content showcasing your expertise within subject areas. Use traditional E-E-A-T (Experience, Expertise, Authority and Trustworthiness) principles to support this. 

• Use statistics and citations to provide evidence and build trust with your audience and generative search engines. 

Outlooks for the future of GEO 

As gen-AI is integrated into more and more search engines, it’s important that your content is optimised to continue to rank highly in the SERP and to appear in the AI-generated summary. Therefore, it is essential to include GEO as part of your ongoing website optimisation strategy. 

A major challenge for generative engines will be developing their ability to retrieve new data from the internet and ‘forget’ out-of-date information, ensuring that summaries given to users are accurate and beneficial (Xiong et al., 2024) and therefore not subject to ‘hallucinating’.  

Going forward, the importance of accurate and precise communication within the life sciences and healthcare communities remains imperative – both to protect scientific integrity and to ensure consumers can make informed decisions. Until generative search engines can fully understand the nuances behind text, data and images (Chauhan, 2024), human input will remain crucial to ensuring content is accurate, up to date and relevant for its audience. 

References

Aggarwal, P., Vishvak Murahari, Tanmay Rajpurohit, Kalyan, A., Narasimhan, K. and Deshpande, A. (2024). GEO: Generative Engine Optimization. KDD ’24: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 33, pp.5–16. doi:https://doi.org/10.1145/3637528.3671900. 

Alkaissi, H. and McFarlane, S. (2023). Artificial hallucinations in ChatGPT: Implications in scientific writing. Cureus, 15(2). doi:https://doi.org/10.7759/cureus.35179. 

Chauhan, C. (2024). The Impact of Generative Artificial Intelligence in Scientific Content Synthesis for Authors. American Journal of Pathology, 194(8). doi:https://doi.org/10.1016/j.ajpath.2024.06.002. 

Duarte, F. (2024). Number of ChatGPT Users (2023). [online] Exploding Topics. Available at: https://explodingtopics.com/blog/chatgpt-users. 

Patel, N. (2024). How to Rank Your Website on ChatGPTHow to Rank Your Website on ChatGPT. [online] NP Digital. Available at: https://neilpatel.com/blog/how-to-rank-your-website-on-chatgpt/. 

Radić, D. (2022). SEO Statistics All Marketers Should Know in 2024 | SerpWatch. [online] Serpwatch. Available at: https://serpwatch.io/blog/seo-statistics/ [Accessed 16 Oct. 2024]. 

Xiong, H., Bian, J., Li, Y., Li, X., Du, M., Wang, S. and Helal, S. (2024). When Search Engine Services meet Large Language Models: Visions and Challenges. [online] Available at: https://arxiv.org/pdf/2407.00128 [Accessed 18 Sep. 2024].