Your Next Customer May Send an Agent Before They Send a Click
For most of the internet’s commercial history, businesses have been marketing to people.
A person had a need.
They searched.
They looked at results.
They visited websites.
They compared options.
Eventually, they chose.
AI has already disrupted parts of that sequence. Buyers can now ask an AI system to research a category, explain alternatives, compare companies, recommend products, or build a shortlist before they ever visit a website.
But the next shift goes further.
Increasingly, the system answering the question will not just help the customer research.
It may research for them.
It may monitor options.
Check availability.
Compare products or providers against specific requirements.
Call a business.
Track a changing condition.
Build a cart.
And, in some cases, complete a transaction.
Your next customer may not arrive alone.
They may send an agent first.
The customer journey is gaining a new participant
This is no longer a distant prediction about where AI might eventually go.
In January 2026, Google introduced the Universal Commerce Protocol, an open standard designed to let AI agents interact with businesses across discovery, buying, and post-purchase experiences.[1]
Google describes the underlying change directly: agentic commerce is an environment “where AI completes tasks on people’s behalf.”
A few months later, Google expanded the concept well beyond a traditional product search.
At Google I/O 2026, Search introduced information agents capable of continuously monitoring the web for a user, expanded agentic booking for local services and experiences, and announced functionality that allows Google to call certain businesses on a customer’s behalf.[2]
Google is also building agentic purchasing into its shopping infrastructure. Its Universal Cart can reason across products from multiple merchants, monitor prices, flag compatibility problems, suggest alternatives, and connect into checkout experiences powered by the Universal Commerce Protocol.[3]
Shopify is moving in the same direction. Its agentic commerce infrastructure is designed to let AI agents discover merchant capabilities, interact with checkout systems, and complete transactions while merchants retain control of the underlying business logic.[4]
This matters because the customer journey is beginning to change in kind, not simply in interface.
Search changed how people found information.
AI answers changed how people consumed information.
Agents can change who performs the work of evaluating that information.
Search introduced you to the customer. An agent may decide whether you deserve the introduction.
Consider the traditional journey for someone choosing a new software platform.
They might search:
“best project management software for small businesses.”
They receive a page of results.
They open several websites.
They read reviews.
They compare pricing.
They request demos.
The companies competing for that customer have multiple opportunities to influence the decision.
Now consider the same need expressed to an agent:
“I need a project management platform for a 25-person consulting firm. We use Google Workspace, need client collaboration, don’t have an IT team, and want to stay under $500 a month. Research the best options and give me the three I should evaluate.”
The customer has expressed far more than a keyword.
They have provided context.
Constraints.
Budget.
Company size.
Technology requirements.
Use case.
And an instruction to narrow the market.
The agent’s job is no longer simply to find relevant webpages.
Its job is to decide what deserves consideration.
That is a fundamentally different visibility problem.
We have already argued that the customer journey increasingly begins before the click and that businesses need to understand whether they are being included, understood, compared, and recommended inside AI-generated answers.
Agents extend that idea one step further.
The pre-click layer can now become a pre-selection layer.
The invisible shortlist becomes even more important
Traditional digital marketing assumes that businesses can observe at least part of the consideration process.
An impression.
A ranking.
A click.
A website visit.
A product view.
A form submission.
Even when attribution is imperfect, there is usually some observable interaction.
Agentic discovery creates the possibility of a different sequence:
The customer defines the need.
The agent researches 20 companies.
It excludes 15.
It investigates five.
It recommends three.
The customer visits two.
Your company was company number six.
Nothing happened in your analytics.
No lost impression.
No abandoned session.
No failed conversion.
No obvious signal that you were considered and eliminated.
You simply never entered the customer’s visible journey.
This is an extension of a problem we explored in Your Website Has Never Mattered More. Your Analytics Have Never Told You Less.
The business sees what happens when a customer reaches it.
But increasingly, some of the most consequential decisions may occur before that point.
An agent makes that measurement gap larger.
Your website now has another audience
This does not make websites less important.
It may make them more important for a different reason.
We previously described the modern website as source infrastructure: the information layer that helps search engines, answer engines, customers, reviewers, journalists, partners, and AI systems understand a business.
Agentic systems raise the stakes of that role.
A human visitor can make sense of ambiguity.
They can interpret a clever headline.
Explore the navigation.
Infer what a company probably does.
Open a PDF.
Call sales.
Ask a follow-up question.
An automated system needs usable evidence.
What does the company offer?
Who is it for?
Where is it available?
What does it cost?
What are the limitations?
How does it compare?
What integrations are supported?
What proof supports its claims?
What locations does it serve?
Is the information current?
Can the system distinguish the company’s own claims from independent evidence?
In commerce, some of this information is becoming explicitly machine-readable. Google’s UCP architecture allows agents to discover merchant capabilities through standardized interfaces, while Google is adding richer Merchant Center attributes intended specifically to improve discovery in conversational commerce.[1]
But the strategic principle extends far beyond ecommerce.
The easier your business is to understand, verify, compare, and evaluate, the easier it becomes for both humans and machines to determine where it belongs.
This is bigger than “optimizing for AI”
There is an understandable temptation to turn every change in AI discovery into another optimization checklist.
Add schema.
Create FAQs.
Publish comparison pages.
Rewrite product descriptions.
Make the site easier to crawl.
Those things can matter.
But they are tactics.
The larger question is strategic:
If an AI system were asked to evaluate your business on behalf of a customer, would it have enough evidence to choose you?
That question reaches beyond SEO.
It touches positioning.
Product marketing.
Pricing transparency.
Competitive differentiation.
Reputation.
Reviews.
Documentation.
Structured information.
Brand consistency.
Entity clarity.
Customer proof.
Inventory and availability.
Location information.
And the quality of the underlying website.
A company cannot solve an unclear market position by adding markup around it.
It cannot solve weak differentiation by making weak differentiation more machine-readable.
It cannot expect an agent to infer a compelling reason to choose the business when the business itself has never clearly articulated one.
The AI era does not eliminate brand strategy.
It makes ambiguity more expensive.
Agents will not only shop
Retail provides the clearest examples today because purchasing has obvious steps that can be standardized.
But limiting the conversation to shopping misses the larger change.
Google’s Search agents already illustrate broader use cases: continuously monitoring information, finding matching apartment listings, locating experiences based on detailed requirements, checking pricing and availability, booking services, and contacting businesses on a user’s behalf.[2]
The same basic model can extend into many customer journeys.
A homeowner could ask an agent to identify three HVAC companies that serve their area, have strong reviews, offer emergency service, and can provide an appointment this week.
A procurement team could ask an agent to research cybersecurity vendors that meet a defined set of requirements.
A student could ask an agent to compare universities based on program strength, cost, location, admissions criteria, and career outcomes.
A car buyer could ask an agent to monitor local inventory for a particular vehicle configuration and alert them when pricing reaches a certain point.
An executive could ask an agent to shortlist software platforms, investigate customer sentiment, compare capabilities, and schedule demonstrations.
In every example, the business may be evaluated before the human directly interacts with it.
That is the real shift.
Agentic commerce is only one visible expression of a much larger idea:
AI is moving from answering questions to participating in decisions.
The new competition may happen before you know you are competing
This changes how businesses should think about competitors, too.
Traditional competitive intelligence usually begins with a known competitive set.
Who ranks against us?
Who bids against us?
Who sells similar products?
Who does our sales team encounter?
But AI systems can create their own interpretation of the market.
They can introduce companies a business does not traditionally consider a competitor.
They can emphasize different differentiators.
They can decide that another business fits the customer’s requirements better.
They can compare categories that companies internally consider separate.
And increasingly, an agent may perform that comparison without generating a visible search session for any of the companies involved.
This is why the argument in The AI Era Won’t Kill Search. It Will Kill Single-Channel Thinking. is becoming more important.
Search rankings still matter.
Traffic still matters.
Advertising still matters.
Conversion still matters.
But they describe only parts of a broader decision environment.
The emerging question is not merely:
Where do we rank?
It is:
When machines help customers decide who belongs in the market, do we survive the selection process?
What businesses should do now
The answer is not to build a strategy around autonomous agents replacing every customer tomorrow.
That would overstate where the technology is today.
Human judgment will remain important. Different industries will adopt agentic behavior at different speeds. Many transactions will continue to require human review, interaction, negotiation, or approval.
The mistake would be assuming that means businesses can ignore the direction of travel.
Google is building agents directly into Search.
Commerce platforms are creating standards that let agents interact with businesses.
AI systems are becoming better at maintaining context, reasoning across requirements, using tools, accessing current information, and taking permitted actions.
The infrastructure for machine-mediated customer journeys is being built now.
Businesses should begin asking:
Can AI systems clearly understand what we offer?
Do they associate us with the categories where we actually compete?
Can they identify who our products or services are best suited for?
Are our differentiators explicit or dependent on human interpretation?
Is pricing, availability, location, product, service, and policy information clear and current?
Can important claims be verified?
Are our strongest proof points accessible?
How are competitors represented when buyers ask comparison and recommendation questions?
Are AI systems selecting competitors we do not currently monitor?
Where are we disappearing before the customer reaches us?
These are increasingly questions of market intelligence, not simply website optimization.
That is the problem Toren Competitive is designed to examine: how AI systems include, describe, compare, and recommend a brand; which competitors are being selected; and what current assets and entity signals may be shaping those outcomes.
The point is not to optimize for one chatbot.
It is to understand a customer journey in which AI systems increasingly participate.
The website was built for visitors. The next era requires it to work for agents too.
For decades, businesses have spent enormous amounts of time asking how to persuade the person who lands on the website.
That question remains important.
But another question is emerging before it:
What caused the customer to arrive in the first place?
Increasingly, the answer may include an AI system that researched the category, interpreted the available evidence, compared the alternatives, and decided that your company belonged on the shortlist.
Or decided that it didn’t.
The click is not disappearing.
The human customer is not disappearing.
Search is not disappearing.
But a new participant is entering the journey between intent and action.
And that participant may increasingly determine which businesses ever get the opportunity to make their case.
For twenty years, websites were built primarily to help businesses persuade humans.
The next era requires something more.
Businesses will need to be clear enough for machines to understand, credible enough for machines to verify, and relevant enough for machines to decide that a human should see them at all.
Sources
[1] Google, January 11, 2026 — New tech and tools for retailers to succeed in an agentic shopping era. Google introduced the Universal Commerce Protocol as an open standard spanning discovery, buying, and post-purchase support and described agentic commerce as AI completing tasks on people’s behalf.
[2] Google, May 19, 2026 — A new era for AI Search. Google announced Search agents, expanded agentic booking, and functionality allowing Search to call certain businesses on users’ behalf.
[3] Google, May 19, 2026 — Introducing the Universal Cart and more ways to help you shop. Google’s Universal Cart uses AI reasoning across merchants and products and connects with UCP-powered checkout experiences.
[4] Shopify, January 11, 2026 — The agentic commerce platform: Shopify connects any merchant to every AI conversation and Building the Universal Commerce Protocol. Shopify describes UCP as infrastructure enabling agents to discover capabilities, interact with merchants, and complete transactions while merchants retain business logic and merchant-of-record responsibilities.
