Preparing for the open agentic web: What brands need to know

The web isn’t what it used to be. Take Google’s results page: position #1 is now often claimed by AI Overviews, meaning brands’ efforts to rank organically or through paid links are less rewarded than before. Moreover, search is no longer confined to traditional engines. Social platforms have long been search destinations, and now AI-powered answer engines like ChatGPT and Perplexity are capturing increasing attention.

“We are witnessing the rise of dual-audience websites,” says Guill Rodas, chief technology officer, APAC, RGA. “Until now, human-centred design dominated digital experiences. Today, AI agents are emerging as a major audience of the web. Therefore, design and engineering must evolve to serve both humans and machines, enabling both to consume information effectively.”

This shift marks the next wave of web design, following the massive mobile revolution. Traditional principles, focused on human readability, SEO for traffic, and click-through rates, are being challenged. AI Overviews present information directly within search results, reducing the need for users to visit websites and this is impacting key metrics like traffic and CTRs.

To adapt, websites must be redesigned with enhanced modularity and structured data, facilitating seamless consumption by AI agents and tools.

“As brands design for both humans and AI agents, website design is evolving from static, visually driven layouts to dynamic, user-centric, query-driven interfaces prioritising accessibility and discoverability,” explains Lars Maehler, client lead, Publicis Media Hong Kong and head of Digitas Hong Kong. “AI algorithms powering search engines and voice assistants don’t ‘browse’ like humans—they crawl content for relevance, context, and structured data to answer specific queries. This encourages a Q&A-focused content strategy, where websites directly address user intent with clear, concise, and valuable answers.”

 

We are already seeing AI agents and generative search change the game. Websites must now be structured to be easily interpreted, evaluated, and acted upon by machines. This means clearly labelling products, services, pricing, and creating content that speaks to both humans and machines.

“We went through a similar shift 25 years ago when traditional search began. Websites had to evolve from purely aesthetic to crawler-friendly,” says Jim Yu, founder and CEO of BrightEdge. “We’re now entering a similar phase, learning what works for AI agents and automated systems.”

AI-driven personalisation becoming the norm

With AI-driven personalisation increasingly common, brands face the challenge of balancing tailored content delivery with maintaining a consistent brand identity.

“AI can personalise content at scale, but without alignment, brands risk fragmentation,” says Etienne Gautheron, managing director, Jellyfish South Korea. “The solution starts with structure: modular content systems, tone-of-voice guidelines, and clear brand frameworks that AI tools like Pencil can use to adapt messaging without going off-brand. Equally important is internal alignment. The best brands rethink how they create and govern content from the ground up, preventing personalisation from descending into chaos.”

David Klein, co-founder of Orange Line, emphasises strategic alignment: “To keep things cohesive, brands need to lock in their voice, tone, and style so personalised content still feels like the same brand. Real-time data monitoring is crucial to course-correct if personalisation pulls the brand apart.”

AI enhances personalisation by leveraging first-party data from Customer Data Platforms (CDPs) to understand user preferences, behaviours, and cultural context.

“For example, a global fashion retailer might recommend minimalist designs to Gen Z in Japan while suggesting vibrant patterns to millennials in Brazil, all while maintaining its signature aesthetic and messaging,” says Maehler. “The key is blending data-driven insights with human oversight. Analytics provide demographic, purchase history, and real-time behaviour insights to deliver personalised content, like tailored emails or dynamic website banners. Human editors ensure tone aligns with brand voice, avoiding generic or mismatched messaging.”

Adapting SEO and content strategies for AI agents

With platforms like Gemini, Claude, Perplexity, Deepseek, and OpenAI becoming new gatekeepers, brands must understand how these agents work, what sources they cite, and how they interpret brand information. Adapting SEO and content strategies for AI agents and voice assistants presents complex challenges in 2025.

One major shift is from short, keyword-driven searches to conversational, long-tail queries. “For example, a user might type ‘best running shoes’ on a traditional search engine, but a voice query could be ‘What are the best running shoes for marathon training under $100?’” says Maehler. “Brands must optimise for natural-language queries by creating content that reflects how people speak. Structured data like FAQ or how-to schemas is critical for AI agents to parse and surface content.”

The rise of zero-click searches, where AI provides answers directly on the search page, reduces website traffic.

 

“Brands can address this by crafting concise, authoritative content targeting ‘position zero’ (featured snippets),” Maehler explains. “A fitness brand, for example, could create a blog post titled ‘Top 5 Marathon Shoes for Beginners’ with a clear Q&A format to capture voice search results and LLM citations.”

AI bias and attribution pose challenges. Large language models prioritise content from high-authority domains, making it harder for smaller brands to compete.

“Building E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) through quality backlinks, expert-authored content, and partnerships with trusted sites is essential,” says Maehler. “Tracking AI-driven traffic with tools like Google Analytics 4 or emerging platforms offering behavioural query insights from LLMs helps brands refine strategies sustainably.”

Staying ahead requires proactive trend spotting and agile infrastructure. Tools like Google Trends, Exploding Topics, and social media platforms reveal emerging topics, like a surge in 'sustainable running gear' queries, allowing brands to create timely content.

“By blending data-driven insights with creative storytelling, brands can thrive in this fast-evolving environment,” adds Maehler.

Brands will also face decisions about allowing AI to ingest intellectual property for participation in conversations or opting out entirely.

“Referrals from AI will become a new channel to monitor, increasing demand for analytics on brand mentions,” says Antonio Panuccio, head of data and tech at Enigma. “Zero-click search and marketing funnel collapse will accelerate. Thirty-second website sessions may deliver thousand-dollar purchases because AI has already matched the product to the user’s needs and preferences, or completed the sale end-to-end in minutes.”

The era of the open agentic web

Emerging technologies like agentic AI, gen AI, blockchain, and Web3 offer opportunities to create web experiences balancing efficiency, authenticity, and engagement for both AI agents and humans.

Gen AI enables fast, personalised content creation at scale. Agentic AI automates interactions and transactions, allowing websites to respond smarter in real time. Blockchain adds security and transparency, building consumer trust, especially for transactions and data handling. Web 3.0 introduces decentralised, immersive experiences connecting platforms and users seamlessly, improving data integration, authenticity verification, and digital environments.

“In practice, imagine a travel site where an AI agent books a flight using verified blockchain data, while generative AI builds a custom itinerary based on user preferences, all without manual searching. That’s where this is heading,” says Gautheron.

 

 

The challenge lies in integrating these technologies across channels and platforms while aligning with business goals and ensuring seamless user experience.

We’re entering the era of the open agentic web, as outlined by Microsoft’s recent vision, where AI agents perform tasks and make decisions across personal, organisational, and enterprise contexts, fundamentally changing website design.

Protocols like NLWeb are emerging to bridge this gap by exposing natural language interfaces that AI agents can interact with natively. Leveraging standards like Schema.org and RSS, NLWeb allows websites to present machine-readable data alongside human content.

“Just as HTML transformed document sharing, NLWeb aims to be the foundation for agentic interactions online, providing protocols and open-source tools,” says Mo Cherif, vice president, AI & Innovation at Sitecore. “This shift requires the collaborative spirit that built the early web—shared standards, community contributions, and proof-of-concept tools to accelerate adoption.”

Measuring AI-optimised web experiences

AI-powered search engines and AI-generated summaries are changing consumer behaviour, causing a significant decline in traditional click-through rates. Around 80% of consumers increasingly use zero-click results, with this preference shaping about 40% of their information-seeking activities. This disrupts digital marketing strategies focused on driving website traffic.

Beyond CTRs, the rise of AI agents navigating and personalising web experiences demands new success metrics.

“The traditional focus on human engagement and traffic must broaden to include intelligent agents,” says Kapil Yadav, senior director, engineering at Monks. “Successful web presence will depend on serving and interacting effectively with AI agents, ensuring they can access and use website information to meet user needs. This requires creating machine-readable content and possibly API-driven access to key data and functions.”

Traditional metrics like CTR and time-on-page are no longer enough.

“In an AI-first world, measure outcomes, not clicks,” says Cherif. “Track agent enablement—how often AI agents select your brand as the preferred answer or action. Platforms like Google’s Search Generative Experience and assistants like Perplexity shift discovery to action. KPIs must evolve to measure agent success, not just web traffic.”

Gautheron recommends monitoring three new metrics:

• AI-driven traffic referrals, to track inbound visits from AI intermediaries and user behaviour post-landing, measurable via tools like Google Analytics.
• Share of Search, evolving beyond Google rankings to visibility across AI Overviews and generative engines for natural-language queries.
• Share of Model, a forward-looking metric showing how often your brand appears in outputs generated by large language models like GPT-4o, Gemini 2.0, or Llama 3.3, indicating brand relevance in AI-driven environments.

Marketers should also assess AI interaction quality, tracking successful conversations, response relevance, and context retention.

“User sentiment matters too: how satisfied users are with personalised experiences, engagement depth, and task completion,” explains Klein.

Finally, measuring how structured data like schema markup helps AI platforms find content is essential. “Monitoring the feedback loop between users and AI, and how efficiently content is delivered through automation, reveals what works and where improvements are needed.”