Generative algorithms have made a silent revolution: they recompose the rules of digital visibility.
Generative algorithms have made a silent revolution: they recompose the rules of digital visibility. In this new landscape, press releases are no longer intended only for journalists, but artificial intelligences capable of analyzing them, synthesizing them and above all deciding on their dissemination. Faced with this mutation, the evolution of CMS (1) is essential to facilitate the understanding, extraction and restitution of corporate content by AI.
The historical tilting of search engines with generative AI
The shattering arrival of chatbots such as chatgpt and AI assistants integrated into search engines created a rupture comparable to Google’s advent in 1998. According to an oxygen study carried out with French journalists, 35.8 % of them now use generative AI tools in their documentary research. A figure that masks a deeper reality: 82 % of press articles citing companies are now written after a prior data consultation provided by AI.
Beyond the production of content, journalists use AI for the analysis of massive data and use it for factual verification. In this context, newsrooms (or press spaces) become strategic interfaces between companies and artificial intelligence.
This mutation comes to challenge the traditional approach to SEO. AIO (Artificial Intelligence Optimization) requires new rules: content structuring, semantic richness, factual data, technical compatibility with AI language models.
When the AIO of a newsroom becomes a performance lever
Of course, the quality of the content is essential, but the technical architecture of a newsroom is just as much. Today, she must speak the language of AI.
AIO performance is based on three main technological pillars:
Advanced semantic markup, for example the automatic integration of metadata (2) enriched, allowing AI to contextualize the information.
Second pillar, certification of the authenticity of sources. Each content generates an unchanging digital imprint, guaranteeing its authenticity in the face of the risk of disinformation. Today a third of CAC 40 information is certified with Wiztrust.
Third pillar, the adaptation of texts to vectorization models (3) of large languages (LLM), very significantly improving their citation rate.
Sorry for the technical jargon but now is the time to get started, it should soon invade the daily life of experienced communicators. It is also the daily life of my teams that continuously optimize our CMS so that the contents of Engie, Crédit Agricole, Disney or Boncoin Source effectively AI.
Beyond traditional contents such as press releases, it is also the profiles of leaders that it is essential to adapt to AI. A hallucination (4) on the CV of the CEO of the company will undoubtedly have an unexpected impact on the reputation of the latter and … his mood of the moment.
Today when a media is looking for information on the pedigree of a manager, our goal is to place his business’s newsroom in the first sources referenced by AI. This is often the first source cited, before its LinkedIn profile.
And when a company presents its results or makes an important announcement, the “winning mix” is the combination of optimization for Google News and generative AI. Immediate visibility via Google News and the sustainability of AI thanks to intelligent archiving compatible with continuous training of linguistic models.
Conclusion: The big return of the RPs and the bottom to the AIO era.
If AI recomposes the rules of visibility, it does not replace the need for quality content, it amplifies it. Human expertise and technological robustness form a virtuous circle. AI also allows you to personalize communication, which should improve the commitment of audiences.
To the communicators who fear this transfer, let us recall the lesson of Gutenberg: each media revolution rewards those who know how to tame the codes. Companies now have the printing of the 21st century – a technology that transforms each word into a credibility atom, readable by humans, search engines and artificial intelligences.
(1) A CMS (Content Management System) is a content management and publication solution. Wiztrust is a CMS intended to optimize the publication of information content on business newsrooms.
(2) Schema.org metadata is structured information added to the HTML code of a web page to precisely describe the content of this page to search engines and other automated systems. This markup transforms the raw data of a site into structured data, allowing indexing robots to better understand the meaning and structure of the information presented
(3) Vectorization is the process which converts gross data into digital representations in the form of vectors, allowing their mathematical processing by computer algorithms. A vectorization model is the set of techniques or algorithms used to carry out this transformation, adapted to the nature of the data (text, image, audio, graph, etc.) and the objectives of the analysis
(4) A hallucination of AI is a generation error where AI invents or distorts facts, while giving the impression of being reliable and precise
