AI is redefining the path from discovery to purchase. As customers shift from typing queries into search engines to prompting AI systems for instant answers, the traditional funnel is collapsing. Agents are beginning to recommend, decide, and even transact on behalf of customers. For brands, this is both an existential risk and a historic opportunity as visibility will no longer hinge on search rankings but on whether an AI includes your brand in its recommendations. To compete, businesses must adapt their data, infrastructure, and partnerships to thrive in a world where the first and last customer touchpoint may be an AI agent.
For decades, search has been the front door to commerce. People typed queries into Google, scrolled through results, and clicked on the links that seemed most relevant. Companies built entire strategies around making sure they ranked higher than their competitors.
That playbook is changing.
Artificial intelligence is reshaping how people find information and how they buy. We are now entering a new phase where customers prompt and receive answers that feel definitive. This shift does more than influence marketing; it is rewriting the rules of customer engagement, brand building, and commerce itself.
So the real question for senior leaders is: How should brands react to these changes in search and online buying behaviour?
We have been experimenting with search and have seen some success with our brand being referenced in AI overviews for certain queries. These early wins are encouraging, but they also remind us how quickly the ground shifts and why experimentation must remain constant.
The New Gatekeepers of Customer Attention
Traditionally, search engines acted as signposts. They pointed users towards websites, where brands had the opportunity to tell their story, showcase products, and build loyalty.
With AI-driven search - through ChatGPT, Perplexity, Gemini, or Claude - the signpost has started turning into the destination. Instead of a list of links, customers now get a neatly packaged answer. It feels authoritative, even if it might not always be accurate.
Imagine asking, “What’s the best smartwatch for runners under $200?” Ten years ago, you would scroll through reviews and sites to compare. Today, an AI agent delivers a shortlist with buying options right inside the chat.
For the customer, this is faster and easier. For brands, it raises two uncomfortable questions:
- What if the AI leaves your brand out?
- What if the AI misrepresents your brand?
When an AI system delivers what looks like a definitive answer, that version of reality can quickly become the truth in the eyes of your customer. And unlike traditional search, you may never get the chance to correct it. There is no page two of search results.
GEO: From SEO to Generative Engine Optimisation
SEO had a clear objective: optimise your site so search engines ranked you higher. GEO works differently. Instead of ranking links, the AI generates an answer by drawing on multiple sources and then reshaping them.
The difficulty is that brands cannot easily predict which content the AI will use, how it will balance different perspectives, or whether it will mention the brand at all. As Ralf Gehrig, CXO from WongDoody, notes, attribution often gets reduced to a minor footnote.
That means lead generation through GEO search may be weaker than in traditional SEO.
Take the example of a premium running shoe brand. In SEO, you could climb the rankings with blogs, reviews, and smart use of keywords. In GEO, an AI assistant might simply say, “The best running shoes this season are from Brand X and Brand Y,” leaving your company out, even if your product is superior.
This quiet disintermediation is one of the biggest risks brands must anticipate.
Customer Behaviour Is Already Shifting
Some leaders assume this transformation will take years to unfold. After all, ChatGPT’s market share in search is just 1% today, and SEO still accounts for half of digital marketing traffic.
But early signals suggest otherwise. Gen Z already prefers prompting over searching. Instead of typing into Google, they ask Perplexity, Grok, or Gemini: “What’s the cheapest way to book a holiday?” or “What’s a good recipe with broccoli, chillies, and rice?”
Customer journeys are being redesigned in real time. The familiar pattern of searching and clicking is giving way to prompting and receiving tailored outputs. The brand’s role in shaping that journey has become more unpredictable. This shift is already altering how discovery, choice, and trust are formed. Brands that assume old habits risk losing visibility at the very moment intent is being shaped.
Our own experiments confirm this: being surfaced in AI overviews does drive new exposure, but those placements can shift overnight with a model update. That is why even after every win, we continue to focus on building adaptability into our strategy.
Agents: From Helpers to Shoppers
SemiAnalysis shows how GPT-5 is moving beyond answering questions to acting as a router, deciding when to allocate deeper reasoning or even initiate a purchase.
Think of it as having an assistant who can recommend the right lawyer and also book the first consultation, or suggest a week’s groceries and then place the order. Early signals of this shift are already visible in partnerships between OpenAI and companies such as Stripe, Visa, Booking.com, and Shopify.
This is “agentic commerce,” where the AI agent evaluates, decides, and transacts. Over time, this evolution could collapse the entire purchasing funnel into a single AI-powered interaction.
The revenue model also changes. Instead of paying for sponsored slots on a results page, brands may face platforms that take a commission on completed transactions.
For any product business, this means the real competition is not for a top listing but to be included in the AI’s shortlist of recommended options.
Why Infrastructure Matters
All of this relies on the AI having access to structured, reliable information.
Parallel’s research systems show that infrastructure designed for agents, not human browsing, can outperform both GPT-5 and human researchers on complex reasoning tasks.
For brands, the implication is clear: if your product data is messy, incomplete, or hidden behind walls, the AI agent may pass you over. If your data is clean, accurate, machine-accessible, and machine-readable, you stand a far better chance of being recommended. The difference lies less in traditional optimisation and more in readiness for machine-to-machine interaction. Structured data becomes the currency that determines whether a brand is surfaced or silently skipped.
We will likely see brands begin mapping an ‘AI journey,’ much like the ‘user journey,’ to anticipate how agents will engage with their data at every stage.
What Brands Should Do Now
For senior leaders and decision-makers, here are some clear imperatives:
1. Invest in structured, trustworthy data
Move beyond keywords. Ensure product information, reviews, availability, and pricing are machine-friendly. You can find schema definitions for these in Schema.org. Also, upgrade feeds, add metadata, and make APIs reliable.
2. Prepare for agent-friendly integration
AI agents need to be able to query your systems and transact directly. Build APIs that allow for stock checks, price comparisons, order fulfilment, and contract renewals. In a B2B context, this could mean asking for a sample, demo, or a quote directly through the AI agent.
3. Differentiate by purchase type
Andreessen Horowitz talks about how purchases fall into four groups: impulse (chocolates), routine (detergent), lifestyle (smart watch), and life-changing (house).
- For impulse and routine, focus on speed and convenience.
- For lifestyle, use curated recommendations that highlight brand personality.
- For life-changing purchases, create AI-powered consultation experiences that reduce risk for the buyer.
4. Balance GEO with SEO
Do not drop SEO as it remains crucial for traffic. Instead, treat GEO as a new channel that demands parallel attention. While SEO is still required, two years from now, GEO might be more prevalent.
5. Build early partnerships
Just as brands that joined e-commerce marketplaces early gained a head start, those who integrate with AI platforms like OpenAI, Shopify, or Stripe will secure stronger visibility when agent commerce scales.
The Metrics Will Change
One of the most profound shifts will be in how success is measured.
Metrics like click-through rates and impressions were built for a world where customers visited websites. And while these metrics will still matter, in a GEO-driven world, the meaningful metric becomes: How often does an AI mention my brand in its answers?
Measuring qualified leads will also still matter, but the very definition of ‘qualified’ will evolve. Instead of 100 website visits yielding 2 leads, you may now see just 10 visits generating the same 2 leads, because the traffic itself is pre-qualified, arriving via AI-driven search results.
New attribution models will start to emerge. Did your customer come from a Google search, a Perplexity response, or a Claude recommendation? The ability to track and adapt to this will separate leaders from laggards.
And while new metrics are invented, owning unique datasets and producing AI-friendly content will be vital advantages.
A Predictive Outlook
Looking ahead, leaders should prepare for:
- Platforms that control the agent interface to dominate customer attention.
- Brands with structured, high-quality data to enjoy disproportionate representation.
- New attribution models focused on mentions inside AI responses rather than clicks.
- Trust becoming the deciding factor: unreliable brands will be deprioritised by agents.
- Early adopters of agent integration to enjoy long-term compounding advantages.
Conclusion: From Optimising for Clicks to Competing for Mentions
Search is undergoing a deep transformation. The customer who once typed and clicked now prompts and receives. The AI that once provided help is starting to make purchasing decisions.
For brands, this presents a pivotal challenge. Without adaptation, visibility could vanish. With preparation, there is an opportunity to be the brand that an agent consistently recommends.
For us, the biggest learning has been that you cannot rest on early wins. The pace of change demands that brands pivot quickly, test often, and build resilience into their digital strategy.
This moment calls for clear decisions: upgrade your data, build integrations, and rethink how you design customer engagement. The first and last touchpoint your customer experiences may well be an AI agent. The question is whether that agent will choose you.
References: