The website of the future: functions and characteristics
In 1 year
The speed at which these technologies evolve remains difficult to predict accurately, week after week, and I don't believe the coming months will bring major upheavals in how websites function, even though recent history has already proven me wrong on this type of prediction. The companies that have not yet integrated these tools into their editorial production may gradually fall behind compared to competitors able to publish faster and more regularly on strategic topics.
Regarding the increasingly popular chatbots, I remain perplexed for now, as a poorly implemented tool that is unable to understand a precise request or redirect to a human contact at the right moment can produce the opposite effect and encourage the user to seek an answer from a competitor. I myself give these assistants one try upon arriving on a site, and an unsatisfactory result makes me leave faster than if I had simply navigated on my own.
The personalization of content and navigation paths is becoming increasingly necessary, both for visitors and for the algorithms responsible for guiding them. An e-commerce site that adapts its product recommendations based on a visitor's browsing history, or a B2B site that highlights different client references based on the detected industry sector, aligns with this logic.
What about SEO in about ten months? Despite concerns raised by the rise of AI assistants, natural search (SEO) remains by far the main source of traffic to websites, as confirmed by a TollBit report relayed by Search Engine Land last October, according to which Google sends 831 times more visitors to publishers than all AI systems combined.
In 5 years
AI engines will gradually establish themselves as permanent intermediaries between users and content. This particularly reminds me of Google's zero position, which for several years has been displaying sports results, weather data, and definitions directly on the results page, inevitably reducing the incentive to click on source sites.
Among the most exposed content will be blogs, guides, comparison sites, and FAQs, whose value relies on information that AI engines easily aggregate and present without the user needing to visit the source site. E-commerce sites will likely lose the upstream phase of the purchasing journey, with discovery and comparison increasingly handled by AI assistants, while transactions, customer account management, and returns processing will remain anchored to the merchant's site.
I believe that the very nature of the content offered on websites should evolve, with theprobable emergence of infographics and explanatory videos generated in real-time by AI based on user actions and requests.
In 10 years
Over the next ten years, websites might rarely be a destination that people visit, but rather a data infrastructure queried by AI agents that will either act on the user's behalf, or on their own initiative by anticipating their needs (renewing an expiring subscription, notifying of a stockout from a regular supplier, suggesting a price alternative).
The situation changes significantly for e-commerce if AI agents become capable of browsing, comparing, ordering, and managing returns autonomously on behalf of the user. We might even expect that administrative websites could also be partially bypassed if AI agents are authorized to fill out official forms on behalf of the user.
If, in a decade, you are among the publishers who have invested in the reliability, structuring, and traceability of their information, you will have a real comparative advantage in an environment where the mass of automated publications will make differentiation increasingly difficult to establish.
The SEO-GEO writer/consultant's support: an AI refiner
My current skills across various practices
It may seem paradoxical, but I spend on average more time on each text than before the emergence of these tools, despite increasingly elaborate prompts. This is for the better, as the content I produce today seems generally more relevant in substance, better structured in its progression, and more enjoyable for the final reader to consume.
Human supervision is not overshadowed by the speed of generation, it requires adapting to it.
My proofreading and correction work is structured around several key areas.
First, stylistically, AI far too often reproduces the same phrasing from one sentence to the next, which therefore needs correction.
As a self-proclaimed AI polisher, here are my top 3 pet peeves.
- "allows"Â
- "while"
- "which"Â
It's also important to vary sentence openings to improve readability, by alternating between prepositional phrases, personal pronouns, determiners, adverbs, or participial constructions, and striving for a fairly literary style. I like to include questions to re-engage the reader and add rhythm. I also believe that each sentence should stand on its own, but also fit well within its paragraph. Yet another stylistic element that writers often struggle with is thejuxtaposition effect, which is a series of individually well-constructed sentences that don't flow naturally.
Beyond the form, I elaborate on certain topics depending on the presumed knowledge level of the target audience. This can involve clarifying a concept,providing an example or making a very clear analogy to ensure the reader's understanding.
Theexpansion of the vocabulary is, in my opinion, an integral part of the work, ideally with the help of tool lexicons like Serpmantics or Thot-SEO. However, it's important not to go to the opposite extreme and vary just for the sake of it. Repeating the same verb twice in a five-sentence paragraph is better than using two distinct verbs, one of which is approximate.
I find it hard to imagine finalizing a text without figures, as quantitative data, in my opinion, helps the reader to visualize. But it's best not to rush into it! AI tools sometimes have the unfortunate tendency, when asked to integrate statistics, to produce their own data by cross-referencing sources whose scopes don't truly align. For every piece of numerical data included in a text, I trace it back to the original source, on reputable authoritative sites, before verifying the information. And it's sometimes tedious, especially when the only figures I consider relevant to add are, it seems – and I sometimes never know for sure – only accessible by paying for an article or a subscription to the media offering that article.
It's important to avoid repeating information from one section to another, I believe they occur even more easily when the brief asks the AI to partially or fully integrate a significant number of keywords. How many times have I been annoyed seeing circumlocutions used to rephrase information that has already been provided. This goes without saying, but it requires multiple readings and a thorough understanding of all communicated ideas.
Thoughts on the evolution of my role
I also anticipate more profound changes, some already underway and expected to become more defined over time, including:Â
- the use of AI to produce content more quickly and consistently, subject to human oversight (I still provide and will continue to provide added value);
- an increasing emphasis on content credibility, in an online landscape where automated sources are proliferating and verification becomes a key differentiator;
- even more in-depth work on information structuring;
- content adaptation to diverse interfaces and formats, from voice search to real-time generated summaries;
- an enhanced value of distinctly human skills, particularly the ability to grasp nuances, apply critical thinking, and understand a user's true intentions;
- the imperative to always keep in mind the objectives assigned to each piece of content, whether to inform, convert, build brand awareness, reassure, or advocate for an idea.
Perhaps with the exception of the most specialized, I believe writers will absolutely need to diversify to survive and, in doing so, be able to take on much more varied assignments than those currently allocated to them.
