"Brands must learn to talk to machines"
At the ComImpact event at Publicis Zurich, it became clear that AI is rapidly changing search behavior—and brands need to respond urgently. Leila Seith Hassan, a British expert at Digitas, called for a radical reframing of existing marketing strategies.
The new search: faster, deeper, more complex
Leila Seith Hassan, Chief Data Officer of Digitas UK, spoke at the Zurich ComImpact event about a paradigm shift: Large Language Models (LLMs) such as ChatGPT, Claude or Perplexity are fundamentally changing search behavior - and faster than any technology before.
On average, users spend over 13 minutes in a session with LLMs. The applications range from simple food ideas to complex questions about finance or health. Particularly relevant for brands: Users place a high level of trust in LLMs - partly because they perceive the answers to be unbiased. As a result, brands themselves have to become an authoritative source in order for them to take place in these systems at all.
From Google optimization to LLM relevance
Serge Barsotti, Managing Director at Digitas Switzerland, described the current phase as the coexistence of traditional SEO and new LLM strategies. While Google is already testing an AI mode in the US, new players such as Deepseek, Claude and Grok are constantly entering the market. As a result, the search landscape is becoming more fragmented, complex and technology-driven.
One key finding: LLMs are developing into agent-based systems that take over tasks themselves - from table reservations to itinerary planning. This integration is not only changing the customer journey, but also the entire conversion funnel. This makes it all the more important to act strategically now.
The two-step model by Leila Seith Hassan: Measure & Transform
1. measure (Measure)
The first step is to understand the current performance of your own brand in various LLMs:
-
How does the brand appear for certain products, target groups or markets?
-
Which sources are used by the LLMs?
-
What is the sentiment of the responses and how does the brand compare with its competitors?
As this information cannot be collected manually in the necessary breadth, specialized tools are required that analyze and aggregate hundreds to thousands of prompts - across different models.
2. transformation/optimization (Transform)
After the measurement comes the targeted optimization of the brand presence:
-
Brands need to develop content, marketing and communication strategies to improve their image in LLMs.
-
The following applies: consider LLMs as a separate audience. They have different value systems, trust in sources and data models - a one-size-fits-all strategy does not work.
-
The aim is to make your own brand a trustworthy source of relevant content - not through AI-generated content en masse, but through high-quality, specific content from credible sources.
-
External elements such as evaluations, research results or credible partners also strengthen perception.
-
Particularly important: creativity and specificity. Generic content is recognized and filtered by LLMs. Only content that stands out remains relevant.
-
And last but not least: Brands have an ethical responsibility. They should set a good example in a field that has hardly been regulated to date.
Conclusion: The relevance of tomorrow is created today
"What brands do today will determine their presence in future LLMs," says Hassan. Those who ignore the developments run the risk of becoming invisible - not in search rankings, but in the answers that users see as the truth. The ComImpact event made it clear: the future of search is AI-driven - and it starts now.
