DMEXCO Column: The Website as a Model of the Past?

AI sets new rules for the online search

We should get used to the idea that in the foreseeable future only a few people will be willing to look at a website or think about the right search terms.

Why? Because AI apps such as ChatGPT and others can already do a lot and simply feel smoother to use than a Google search. At the same time, search methods such as voice search or visual search are playing an increasingly important role.

Briefly explained: zero-click searches are all searches where it’s no longer necessary to click through to the next page because Google has already summarized all the relevant information in bite-sized form on the search engine results page (SERP). These may be mainly knowledge queries right now, but this overall trend will certainly continue.

What does this mean for marketing?

Important marketing touchpoints will vanish, and the convenience demanded by users when receiving information will continue to increase. In combination with AI language models, this means that the provision of information and the respective online source may become completely detached in some cases.

Thought through, in extreme cases this would mean that websites become pure information repositories from which crawlers and bots are happy to help themselves, but on the other hand generate their own “end product” from it, which they ultimately serve to the user in their own interface.

Applied to the healthcare sector, this would mean that questions such as “I have severe back pain, what can I do about it?” or “Drug XY has given me a rash, what alternatives are there?” are still asked when the person concerned thinks of them. And of course, it makes sense that answers to such questions are given directly.

Now, we all know that a medically correct answer can sometimes be very tricky – not to mention the liability issues that come into play. Instead of giving a medically sound answer, AI developers might therefore be more inclined to give “organizational answers” or further advice, for example: “If you have severe back pain, I recommend you see a specialist. Dr. Antipain’s practice is 800 meters from your current location. Would you like me to make an appointment for you?”

As a user, I would then no longer have to even look at the actual website. This would rapidly reduce the importance of company websites – but also the trust in certain information.

Marketing departments and agencies might ask themselves: “Quo vadis SEA?”, “Quo vadis content?” – and above all, “Quo vadis customer relations?”.

Living in a zero-click world

One thing is certain: Google and OpenAI want to earn money with their products. When an innovative application is introduced, the second step will be to attract as many users as possible, and the next step will be to start monetizing it. In this case, this will mean new forms of advertising being integrated into the newly created user journeys.

For example, slots for “contextual information” about who this content originally came from and that this content is trustworthy. Or the reference to deep dives and interactive content. Although these deep dives are also offered directly in the AI app dialog, they are clearly labeled as advertising.

#1 Make a decision about your own content!

Usually, every healthcare organization has enough valuable content on drugs, studies, therapies, etc. These organizations should decide: Do I want to make all or part of my content “open source” – meaning also available to AI platforms?

Or is my content so exclusive and valuable that I want to protect it from unauthorized access by AI tools of all kinds, for example by a login?

Particularly in the consumer sector (or equivalently in healthcare marketing, for example patient programs or therapy information), however, good arguments are needed as to why content should be protected by a login and not made publicly available. While this is a common way of dealing with information for healthcare professionals and medical communities, and is often even required by the Medicinal Products Act, this must be assessed separately depending on each content.

A balance must be found on a case-by-case basis: Do you want to provide users with good content online and also get involved in the AI game, or do you want to preserve your “content treasure”?

#2 The game goes both ways!

Just as AI models make use of online content for their functionality, content creators can make use of existing AI frameworks.

Owned content can be used to develop owned services such as AI-supported chatbots to offer them to your own target group – or to develop your own “Corporate AI” that uses LLMs (Large Language Models) of the relevant providers and adds or enhances your own specific content. In the long term, refined company-specific AI feeds are very possible.