By Jack Smyth, country lead, The Brandtech Group, Jellyfish, Australia
Marketing is all about eras. The 2000s were the ‘direct’ era – when digital publishers, brands, and retailers were focused on driving people to their owned platforms.
In the 2010s, we had the ‘distributed’ era, catalysed by social and mobile, when brands met us where we were, on social feeds and platforms, and the KPIs of likes, shares, and comments became the currency by which success was measured.
But still the target audience was human. Even the algorithms were designed to influence human behaviour, because ultimately, it was humans who saw, clicked and converted.
The current era challenges this idea, and, in some ways, the future of brand strategy itself. This is the agentic era – a time when brands, publishers and retailers need to reassess their strategic approaches.
This time it’s different: the AI shift
Brands have historically been made by people, for people. Today, though, there’s a new kid on the block: AI. It plays multiple roles: as a gatekeeper of information; as an audience that marketers need to consider; and as a buyer, making purchases with or on behalf of people.
In the agentic era, brands need to adapt to the prominence of AI at every stage of the consumer journey, updating their processes and outputs to build brands and businesses that succeed in a dramatically different context.
To date, much of the focus on AI and LLMs has been on generative engine optimisation (GEO), which responds to changes in the way we make text queries. But there is more going on that should excite marketers:
- Agentic shopping, such as ChatGPT Agent, has the potential to disrupt digital commerce through new methods of transaction and paths to purchase.
- The arrival of paid advertising inside AI interfaces. This is a game-changer. AI interfaces are beginning to monetise their surfaces, and the brands that understand how to buy visibility inside a reasoning engine will have a significant early mover advantage. This isn’t traditional display. It’s about understanding how a sponsored result earns trust inside a conversational answer.
- AI browsers, such as Perplexity Comet, create new ways to engage with AI, integrating their capabilities into every digital impression.
- Multimodal search, such as Gemini Stream, allows people to bridge the gap between the physical environment and AI by using image and video inputs to ‘talk’ directly to LLMs.
A modern marketing system
Given all these shifts, to succeed in the agentic era, brands clearly need to make themselves visible to the machines. The approach we propose is Generative Engine Marketing (GEM).
GEM represents a systematic framework to ensure that brands are accurately understood and effectively surfaced by AI systems. Building on the Jellyfish Share of Model™ platform – which enables companies to analyse how LLMs perceive their brands, products, and services – GEM provides a continuous loop of technical optimisation, content creation, pre-testing, distribution, and measurement, to ensure that brand assets are interpretable by models, resonant to audiences (both human and agentic), and optimised through ongoing feedback.
Share of Model™ enables marketers to see how brands actually appear inside the models shaping the next era of discovery. It is built around four modules – brand perception, search visibility, asset evaluation, and activation – that help brands build their GEM strategy.
These allow us to measure salient metrics such as mention rate and average position, showing how often and how prominently a brand appears in model responses. We can map category perceptions, revealing how different models describe a brand’s strengths, weaknesses, and associations relative to competitors. Sources of influence can be traced to understand where models themselves look for information and answers. And it’s now on the brand’s own media. We’re taking influencer content, community content from sites such as Reddit, and niche publishers focused on every hobby you can think of, from running to stamp collecting.
From diagnosis to design
Once this landscape is visible to us, we can move from diagnosis to design – optimising the technical layers that determine visibility and authority. Every piece of brand communication becomes a signal that informs how AI understands and recommends it. The brands that learn to optimise for this layer will not just be found, but preferred by human and agentic audiences.
The results of activating LLM insights into paid media can be material and represent an immediate return on investment for GEM.
Gentle Monster, the Korean eyewear brand, used LLM insights to optimise its Google Performance Max campaigns before the US holiday season, generating a 17 per cent improvement in click-through rate, a 14 per cent uplift in conversion rate, and a 39 per cent improvement in ROAS.
Industrial maintenance, repair, and operations company, MSC Industrial, drove a 45 per cent increase in revenue and a 758 per cent incremental ROAS in the first 30 days of implementing AI-driven optimisations to its Performance Max campaigns.
To operationalise a GEM strategy, brands must translate insight into action through five interconnected levers: technical optimisation, content creation, pre-testing, distribution and measurement.
These levers play into different moments of the marketing lifecycle:
- Launching a new product: ensure LLMs can discover and describe your product accurately before launch by optimising its structured data, brand narratives, and category associations across text, image, and video assets
- Improving a brand’s position: analyse how LLMs perceive your brand versus competitors, and then use content creation and distribution levers to reinforce or reframe key attributes
- Entering a new market: pre-test assets with AI models to identify linguistic, cultural and semantic nuances, and track how model visibility and perception shift as localised content is deployed.
Marketing has always been about managing change.. Each shift in media technology has redrawn the boundaries of how brands communicate, compete and resonate. But this time, the shift is different, bigger.
GEM represents a challenge, but also, an opportunity: to design brands that are visible to the LLMs, resonant to humans and optimised for both. It extends the marketer’s craft beyond campaigns into the underlying data, structures and signals that inform AI understanding. In this sense, GEM is less a new tactic, more a new way of marketing.
AI does not end the story of marketing, but it does open a new chapter. GEM provides a canvas on which it can be crafted.
