This is how you bring your brands into ChatGPT and Co
Generative AI Optimization (GAIO) is a hypothetical marketing discipline, analogous to Search Engine Optimization (SEO). One can also speak of chatbot optimization, which is used by Philip Kloeckner However, the proposed term GAIO also makes sense.
GAIO is about how companies can ensure that their brand and their products appear as prominently as possible in the widespread Large Language Models (LLM). Because models such as ChatGPT, Bing Chat, Google Bard or Google’s Search Generative Experience could influence many purchasing decisions in the future – and are probably already doing so to an increasing extent today.
For example, if you ask Bing Chat about the best toaster for a shared kitchen, the AI will recommend a Philips device. And anyone who then pretends to be the managing director of a growing small business and asks for help in selecting suitable ERP software will be recommended SAP Business One. How do these answers come about?
Bing Chat’s recommendations always take into account the context provided. A gamer, for example, is recommended a different laptop than a graphic designer. The AI refers to secondary sources such as specialist media, blogs or comparison sites from which such recommendations are fed. Manufacturer sites, on the other hand, almost never use them as a source.
Editor’s Recommendations
The established basic rules of search engine optimization are thus undermined: In the classic web search, backlinks are still the most important relevance criterion. Those who can collect more high-quality links have a clear advantage in the fight for the first places in the search results.
The chat-based search seems to behave differently: Here it is no longer about backlinks, but about brand mentions. It is no longer important where I am linked, but where I am mentioned and what is said about me there – good news for PR consultants. At the same time, your own website loses importance if search engine users are redirected there less frequently.
At the beginning of this text GAIO was described as a “hypothetical” marketing discipline, perhaps one could also speak of an “emerging” discipline. There are two hurdles to overcome: First, a critical mass of users must be convinced of chat-based search. Secondly, the question of what the GAIO method case could look like needs to be clarified.
In order to win and keep enough users, the AI simply has to get better first – especially when it comes to product recommendations and product-related questions. While Bing Chat is already quite good at issuing context-related and even comprehensibly justified recommendations, Google Bard currently only provides short product lists without justification and source information, which offer the user little added value.
ChatGPT, on the other hand, offers an enormous amount of advice and context, but always strikes when it comes to recommending specific products. This is because the training data for this LLM currently only extends to September 2021 and unlike Bard and Bing Chat, the AI cannot yet access live data from the Internet. However, if the further development of the technology continues at a similar speed as in the past few months, such problems should soon be resolved.
How will GAIO work in the future? The most immediate problem for companies is certainly the lack of transparency: they currently do not know how present their products and brands actually are in the answers of the LLM. Which questions are asked how often? Do competitors possibly do much better in the answers? Is incorrect or outdated information circulating? In which contexts are which products recommended – and from which sources do these recommendations come from? In order to answer such questions, new tools are required that will work in a very similar way to the widespread analysis tools in the SEO field, such as Semrush or Sistrix.
Once this transparency has been created, prospective GAIO specialists must then learn to understand which factors influence the presence of a company in the answers of the LLM. So it’s about those adjustment screws that are referred to as ranking factors in search engine optimization.
So far, Bing Chat, which links information sources under its answers, has provided the most clues as to how answers came about. At least ChatGPT knows which training data the model underlie and how these are weighted – a good starting point for further exploration.
We can only start turning them once we know what the adjustment screws are. Then the optimization really starts. A relatively simple tactic could be to systematically evaluate which websites consistently source the LLM in certain product categories. You can then try to place your own products and messages on these pages in a targeted manner.
In order to be able to understand whether such measures actually make a difference, a success measurement is finally necessary. So far, all the prerequisites for this have not been met, because the LLM providers have not yet provided any figures. But that should change in the next few weeks: Microsoft wants it take the first stepby integrating Bing Chat Reports into its webmaster tools.
In addition, Bing Chat should be specifically shown in the referrer data so that site operators can understand how many visitors the chatbot directs to their sites. From this point on, it will be possible to draw first conclusions and to better understand the relevance of Generative AI in the customer journey.
Is GAIO more than a thought experiment? Yes – as long as Generative AI can permanently establish itself as a research tool. By then, at the latest, companies will be asking themselves how they can increase the likelihood of appearing in the answers of the LLM. But as with most fast-growing digital channels, there’s a significant first-mover advantage. It could therefore be worth dealing with the topic early on.
