When migrating an e-commerce site to a new CMS in 2024, one quickly encounters technical choices that didn’t exist two years earlier. Should we integrate an internal search engine powered by generative AI? Adopt passkeys for customer authentication? These questions influence the budget, timeline, and technical stack of each web project.
Passkeys and passwordless authentication: what it changes on the front end
In a recent project to redesign the customer area, we replaced the username/password combination with WebAuthn-compatible passkeys. The user benefit is immediate: no more password reset forms, no more friction at login. Resistance to phishing is also significantly better, as the cryptographic key is tied to the domain.
The WebAuthn Level 3 specification now allows a single passkey to cover multiple related domains. For brands managing multiple sites (store, blog, customer service area), this is a concrete operational change. Authentication is configured once, and it works across the entire ecosystem.
Some teams report integration difficulties with less common browsers. Attacks targeting passkeys synchronized via Google on Windows have also been documented. Adopting passkeys does not eliminate all attack surfaces, but shifts the risk to device management rather than user memory.
For companies, regularly publishing tech news on The Web Brains’ website helps keep track of these developments over time.
European AI Act: the regulatory timeline that structures web projects
The European AI Act came into effect on August 1, 2024. This is not a theoretical text: it applies in successive phases, and each phase modifies what can be deployed on a site or application.
Since February 2025, certain AI practices are prohibited (social scoring, behavioral manipulation). From August 2025, obligations will apply to general-purpose AI models (GPAI), including those integrated via third-party APIs for automated content or product recommendations. Transparency obligations will begin in August 2026.
Specifically, if we use an AI chatbot on an e-commerce site or a content generator to feed a blog, we need to ensure that the model provider complies with GPAI obligations. Compliance no longer rests solely on the site publisher but on the entire technical subcontracting chain.
Points to check before integrating an AI tool on your site
- Has the model provider published technical documentation that complies with the GPAI requirements of the AI Act, particularly regarding training data and systemic risks?
- Does the user interface clearly indicate that the content is generated by artificial intelligence, as required by the transparency phase?
- Do the personal data collected by the tool comply with GDPR in addition to the AI Act, with an up-to-date DPA (Data Processing Agreement)?
SEO strategy and AI-generated content: field trade-offs
Google does not prohibit AI-generated content, but its recent updates penalize content without added value, whether human or automated. In practice, we observe that mixed pages (AI structure + editorial rewriting) perform better than fully automated or fully manual content that is slow to produce.
The challenge is not to choose between AI and a writer. It’s about defining a workflow where automation speeds up production without sacrificing relevance. A structured brief, a first draft generation, followed by a human review that adds field data, concrete examples, and relevant internal links.
Content marketing automation tools
Tools that combine text generation and SEO optimization have multiplied. The classic pitfall: mass-producing content that all looks the same and dilutes domain authority. The winning strategy remains to publish less but with a higher informational density, targeting specific queries rather than generic keywords.
For brands managing their audience on social media in parallel, the same principle applies. Engagement does not come from the volume of posts but from the relevance of the content to what the audience is actually seeking. The influencers who perform well in 2024 are those who produce specialized content, not those who post daily without an editorial line.
Web performance and Core Web Vitals: the thresholds that really matter
We often spend too much time optimizing the Largest Contentful Paint while the real issue lies with the Interaction to Next Paint (INP), which replaced First Input Delay in March 2024. This metric change alters technical priorities.
INP measures the overall responsiveness of the page throughout the session, not just at the first click. Sites that load heavy third-party scripts (analytics, A/B testing, online chat) are the first to be penalized. Before adding a new marketing tool to the site, we test its impact on INP under real conditions, not just in a lab.
- Audit each third-party script with Chrome’s DevTools (Performance tab) to measure its impact on the main thread
- Defer loading non-critical widgets (chatbot, newsletter pop-up) after the user’s initial interaction
- Prioritize server-side solutions for personalization rather than client-side scripts that block rendering
The relationship between web performance and conversion rate remains direct. A fast and responsive site converts better, retains users longer, and sends better signals to Google. The trend for 2024 is not to add features, but to remove what slows down without providing measurable value.
These four axes (authentication, AI regulation, content strategy, technical performance) are not decorative trends. They are the topics on which web teams spend time each week, with real budget trade-offs and measurable impacts on traffic, security, and compliance.
