Cloudflare CEO Matthew Prince warned during an Axios interview at Cannes Lions that AI agents could make small businesses harder for customers to find, compare, and choose as search and commerce move from human browsing to AI-generated answers and recommendations.

Prince’s concern goes beyond publishers losing traffic to AI summaries. His larger argument is that customers are increasingly relying on AI agents to research products, compare vendors, and decide what to buy and where to buy it. That behavior could reward larger, established companies with stronger reviews, richer data, broader citations, clearer policies, and larger marketing footprints.

For small businesses, the main decision is whether to wait and see how AI-driven discovery develops or start building the public trust signals AI systems may use to find, compare, and recommend them.

The issue is already visible in media, where AI answer engines can summarize original reporting and reduce click-through traffic to publishers. The same discovery pressure could eventually affect retailers, software companies, professional services, agencies, local businesses, and new consumer brands.

In short, AI may give small businesses powerful operating tools while making customer discovery more dependent on the trust signals AI systems can find, compare, and recommend.

Agentic commerce refers to the use of AI agents to research, compare, recommend, or complete purchases on behalf of consumers or businesses.

AI-driven discovery is the shift from customers browsing search results to AI systems summarizing, ranking, or recommending the businesses customers see first.

Search is already changing from a traffic engine into an answer engine.

For years, publishers and businesses built their online strategies around a basic exchange: search engines crawled their pages, ranked their content, and sent users back to the original site. That exchange worked because users still had to click through to read the full article, compare sources, evaluate products, or make a purchase decision.

AI search changes that behavior.

When Google places an AI-generated answer at the top of a search page, or when a chatbot answers a question directly, users may get enough information without visiting the original source. The publisher or business may still be used to generate the answer, but it may not receive the click, the reader, the customer relationship, or the revenue opportunity.

That issue surfaced at the Cannes Lions advertising festival in France, where Axios interviewed Cloudflare CEO Matthew Prince about AI, publishers, bots, and the future of the open internet.

Prince focused first on the pressure AI systems are placing on publishers. He argued that people increasingly trust AI interfaces and are less likely to click through to original sources. That makes it harder for publishers to turn reporting, analysis, and other original content into traffic, advertising revenue, or subscriptions.

According to Axios, Prince described a sharp change in the crawling-to-referral ratio, which compares how often a platform crawls, reads, or uses a publisher’s pages with how often it sends a visitor back to the publisher’s website.

Ten years ago, Prince said Google crawled about two pages for every visitor it sent to a publisher. He now puts that ratio at 18:1 for Google, 1,500:1 for OpenAI, and 60,000:1 for Anthropic.

Those larger numbers mean AI systems and search platforms may be using far more publisher content while sending back far fewer visitors. Prince’s explanation was direct: “People aren’t following the footnotes.” In other words, users may see citations or source links, but they are not clicking those links to visit the original websites.

Publishing shows this problem early because it depends directly on information discovery. A news organization can produce the reporting or analysis that helps answer a question, but if the AI system summarizes that work inside the answer, the reader may never leave the AI interface.

That is why media may be the early warning sign for a much larger business issue.

Prince’s small-business warning centers on customer access. AI may make smaller companies more capable internally, but agentic commerce could create a new gatekeeper if agents become the layer between buyers and sellers.

When AI chooses which sources to cite, publishers compete for inclusion in an answer. When AI agents choose which products, vendors, or services to recommend, businesses may compete for inclusion in a customer’s decision set.

The same mechanism that changes media traffic could eventually change customer acquisition.

Prince’s bigger concern is agentic commerce.

In traditional online shopping, the customer still does much of the discovery work. A person searches, clicks, compares options, reads reviews, evaluates brands, and decides what to buy. That process gives smaller companies several ways to earn attention even when they are not the largest player in the market.

A small business can appear in search results, run targeted ads, show up in social feeds, earn referrals, build a loyal audience, or persuade customers through product quality, service, storytelling, and brand identity. The buyer may compare a large retailer with a niche company and decide the smaller option feels more personal, specialized, or trustworthy.

AI agents could compress that process.

If a consumer asks an AI agent to find a product, book a service, choose a vendor, or recommend a supplier, the agent may evaluate hundreds or thousands of options before the customer sees anything. It may compare price, availability, delivery speed, return policies, customer reviews, reputation, third-party citations, business history, support quality, and structured information across the web.

The result may be a short recommendation, a ranked list, a link to a preferred vendor, or eventually an option to complete the purchase inside the AI experience.

That could be convenient for consumers. It could also make discovery much harder for businesses that do not make the agent’s shortlist.

Prince argued that agents will not respond to brands the same way humans do. Prince connected the issue to the way people choose brands today. Consumers often buy from a company because it is convenient or because they feel some loyalty, familiarity, or trust toward the brand. An AI agent changes both parts of that decision.

Because an agent can compare thousands of options quickly, familiar brands may lose some of the advantage they have with busy human shoppers. It also does not have the same emotional attachment to a logo, store design, packaging, or brand story. For Prince, that raises a new question: what does brand mean when the buyer is an AI agent?

In that market, brand may become less about visual identity and more about evidence. An AI agent may place more weight on measurable trust signals, such as whether the product arrives on time, whether quality is consistent, whether policies are clear, whether customers are satisfied, and whether outside sources support the company’s claims. That changes the competitive challenge.

A business still needs to attract customers, but it may also need to prove to AI systems that it is credible enough to be considered. The company is no longer competing only for a person’s attention. It may also be competing for the agent’s confidence.

For large companies, that discovery challenge may be easier. They often have more reviews, more backlinks, more media coverage, more structured data, more advertising history, more customer records, and more third-party validation across the internet.

For smaller companies, being good may not be enough. They may also need to make their credibility clear enough for AI systems to find, scan, compare, and trust.

One common response to Prince’s warning is that businesses have adapted to major discovery shifts before.

The internet changed how companies were found. Google changed how companies competed for visibility. Social media changed how companies built audiences and customer relationships. Search engine optimization became a normal part of operating online, and many small businesses learned how to compete inside that system.

That history is part of the reason some observers see AI optimization as the next stage of the same pattern. Businesses may adapt again by improving their websites, creating clearer content, earning reviews, using structured data, and building the authority signals AI systems can recognize.

But there is also a key difference.

SEO was built around ranking. A business could appear on a search results page, even if it was not the first result. Ranking fifth, seventh, or tenth still gave a company a chance to be seen, clicked, compared, and considered.

AI-driven discovery does not work the same way.

When an AI system answers a question, it may gather information from websites, news articles, reviews, business listings, product pages, documentation, forums, and other available sources. Instead of sending the user to a page of links, it may combine that information into one answer. In other cases, it may present a short list of recommended products, services, vendors, or sources.

That changes the competitive pressure.

In traditional SEO, the goal was to rank high enough for a customer to find the business. In AI-driven discovery, the goal may be to make the shortlist before the customer ever sees the available choices.

That is why the comparison to SEO is useful but incomplete.

A business that ranks lower in traditional search can still be visible. A business that does not make an AI agent’s shortlist may never reach the customer at all.

That is the harder problem. Search engine optimization, or SEO, helped businesses compete for placement in search results. Answer engine optimization, or AEO, is about making information clear, credible, and structured enough to have a chance of appearing in AI-generated answers.

AI agents add another layer. A business still needs to be found. It also needs to be selected by a system it does not fully control.

If AI agents become major intermediaries between businesses and customers, new entrants may find it harder to break into markets.

That is where Prince moved from publishers to small businesses. If a small business is trying to convince an AI agent to buy from it, he asked, “How do you do that?”

His concern is that the established players could become more powerful if new entrants cannot get into the agent’s recommendation set. Over time, that could lead to more consolidation and fewer companies breaking into markets that are already dominated by large incumbents.

The risk could appear in several ways.

Established companies may benefit from trust signals they already have. Large companies often have years of reviews, press mentions, customer data, partner ecosystems, backlinks, business listings, and brand familiarity. AI systems may treat those signals as evidence of reliability, trust, and authority.

Smaller companies may have fewer external validation points. A new business may offer strong products or services, but if the internet contains limited information about it, fewer backlinks, fewer press mentions, fewer reviews, or inconsistent business details, an AI system may have less evidence to evaluate. Compared with a larger company that has years of public signals, the smaller business may appear less established or less trustworthy.

AI agents may also favor lower-risk recommendations. If an agent is trying to avoid a bad customer experience, it may choose companies with proven fulfillment, clear return policies, strong support records, and widely available reputation data. A smaller business with few reviews or limited public information about customer service may look less reliable, even if it provides a better product or more personal experience.

The number of visible choices may also decline. If an AI agent recommends three vendors instead of showing 30 search results, many businesses may never appear in front of the buyer. To the customer, those businesses may effectively never exist.

That could lead to more concentration around established players, especially in markets where trust, fulfillment, price comparison, and reputation are easy to measure. Over time, that kind of recommendation pattern could push smaller companies further to the margins.

Prince said he is already discussing the issue with companies that serve small businesses, including Shopify and major payment networks. That matters because the problem is not limited to publishers. If agentic commerce grows, online stores, payment companies, marketplaces, search platforms, and business software providers may need new ways to help smaller companies prove they are credible options.

Retail is a clear example. If a customer asks an agent to buy a household item, will the agent choose a smaller brand with limited review history or a large retailer with extensive logistics, return policies, and customer data?

The same concern could apply to software, marketing services, legal services, accounting, consulting, healthcare navigation, travel, and local service providers. In those markets, AI systems may look for evidence of experience, customer satisfaction, professional credibility, response quality, pricing transparency, and third-party validation. Businesses with fewer public signals may have a harder time proving they deserve to be recommended.

The danger is not that every small business disappears. The danger is that customer attention becomes harder to earn before a human even sees the options.

Prince’s warning drew a range of responses on X, including the argument that AI may help small businesses more than it hurts them.

AI gives small teams access to capabilities that once required far more people. A five-person company can now use AI for writing, design, research, coding, data analysis, customer service, administrative work, and marketing support. That can reduce overhead, expand capacity, and allow a solo entrepreneur or small team to produce work that once required agencies, developers, designers, analysts, or support staff.

That is a major operational advantage, but it does not answer the customer acquisition question. A small business can become more efficient, more productive, and more capable with AI while still struggling to be found, trusted, or included in an AI-generated recommendation.

Some critics argued that people will not use AI agents for every purchase. Consumers may still want personal involvement in major decisions, emotional purchases, luxury goods, travel, home design, healthcare, and experiences. But AI does not need to control every purchase to influence markets. If customers use AI to research options, compare vendors, review companies, or build a shortlist, the agent may shape the decision before the customer starts evaluating the full range of choices.

Local and community-based businesses may also be more resilient. Restaurants, repair shops, boutiques, fitness studios, salons, local service providers, and experiential businesses often depend on human relationships, physical proximity, trust, and community reputation. But that resilience has limits. A business that relies only on neighborhood recognition may stay visible to existing customers while missing online sales, tourism traffic, service-area searches, regional growth, or AI-generated local recommendations.

Some commenters argued that AI will compress margins rather than destroy companies. If agents can compare prices, quality, warranties, delivery speeds, and service terms instantly, businesses may face pressure to become cheaper, better, faster, or more specialized.

For smaller companies, that can still create a serious profitability problem. Lower prices and tighter margins leave less room to invest in content, reviews, customer support, technology, fulfillment, and the credibility signals that help AI systems evaluate a business. A company may survive that pressure, but it may have less revenue, less flexibility, and fewer resources to compete.

Others compared the shift to the early internet. Every major platform shift created fear, but new companies still emerged. The internet did not eliminate small businesses. In many cases, it helped create them.

That comparison is useful, but it does not remove the competitive pressure. The internet gave small businesses new ways to publish, sell, and reach customers. It also created new gatekeepers, including search engines, social platforms, marketplaces, and app stores. Businesses adapted, but they also had to learn new rules to compete effectively.

AI may follow the same pattern. It may make small businesses more capable while making discovery more demanding.

The concern is that adapting may not be evenly distributed. Larger companies already have more reviews, more citations, more customer data, more media coverage, and more money to strengthen those signals. Smaller companies may have to work harder just to prove they belong in the same recommendation set.

AI may give small businesses better tools to do more with less, but it may also make the path to customer acquisition and growth narrower, more expensive, and harder to control.

When fewer small businesses can reach customers, markets become less competitive. Consumers see fewer choices, new entrants face higher barriers, and more power concentrates around companies that already have scale.

Media is one of the first industries to feel this shift because AI systems are already good at answering information-based questions.

A user can ask for an explanation, a summary, a news update, or a list of sources, and the AI system can produce an answer using information gathered from across the web. That does not require a checkout process, delivery network, return policy, payment system, or customer service handoff. It only requires the AI system to find, summarize, and present information.

That makes publishing an early test case for AI-driven discovery.

For publishers, the business problem is direct. A news organization may report the story, explain the issue, or provide the context that helps an AI system answer a question. If the reader gets enough information inside the AI interface, the publisher may lose the site visit that supports advertising, subscriptions, newsletter signups, brand recognition, and long-term audience growth.

The impact is serious for large publishers because even established media companies depend on recurring traffic and direct audience relationships. If search and AI systems use their work while sending back fewer readers, the business model becomes harder to sustain.

For smaller publishers, the challenge is sharper. They already have less brand recognition, fewer backlinks, fewer citations, smaller audiences, and less historical authority across the web. AI systems may still find their work, but established media organizations often have more visible trust signals.

AiNews.com has seen that challenge directly through AI discoverability testing. Established media organizations appeared more often in AI-generated recommendations, even when smaller publications were producing relevant or deeper coverage in the same category.

AiNews was not invisible. In one third-party dataset reviewed by AiNews, the site appeared among recommended AI news sources, but lower than more established competitors. That result is important because it shows the problem is not total exclusion. A smaller publication can appear, but it may still receive a smaller share of visibility than larger brands with more public authority signals.

That is the business lesson, not a media-specific complaint.

If AI systems tend to reward the companies with the most visible trust signals, the same pattern could affect any market where customers ask AI for recommendations. Media shows the issue first because information is easier for AI to summarize than a purchase is to complete. Retail, software, professional services, travel, healthcare navigation, and local services may face a similar challenge as AI agents become more involved in research, comparison, and buying decisions.

The concern is not whether every small publisher or small business can survive. Many will. The larger question is whether new and niche companies can still earn enough visibility to grow when AI systems increasingly decide which options customers see first.

The outlook is not entirely negative for smaller companies.

If AI systems improve at evaluating quality, relevance, and trust, smaller businesses may still have a path to visibility. They may not win the broadest searches or the most generic product recommendations as often as larger companies, but they may be able to compete when the customer’s need is more specific.

That could matter for niche products, local services, specialized expertise, faster availability, stronger customer service, unique inventory, or businesses with a clear reputation in a defined category.

The opportunity is not that smaller companies can simply wait for AI systems to notice them. They may need to make their credibility easier to find, verify, and compare.

That includes:

Those signals matter because AI systems may look beyond a company’s own marketing claims. They may compare what the business says about itself with what customers, directories, publishers, platforms, and other outside sources say about it.

Large companies may be recommended more easily because they already have broad recognition and many existing authority signals. Smaller companies may need to create those signals more deliberately by earning reviews, clarifying their expertise, improving their website structure, building direct customer relationships, and being cited by external sources.

That is more difficult than traditional SEO, but it is not impossible.

The realistic opportunity is not equal visibility across the entire market. It is a better chance to appear when the customer is looking for the kind of product, service, location, or expertise the smaller business can credibly provide.

Like traditional SEO, AEO is not a magic bullet. Search engine optimization gave businesses a chance to appear higher in search rankings. Answer engine optimization, or AEO, gives businesses a chance to appear in AI-generated answers and recommendations.

In an agent-driven market, being considered may become the new competitive threshold.

Businesses do not need to wait for agentic commerce to become mainstream before adapting. The earlier a company starts building clear information, reviews, citations, direct customer relationships, and public credibility signals, the more prepared it may be if AI systems become a larger part of customer discovery.

The first step is to make the business easier for customers, search engines, and AI systems to understand, verify, and trust. That means treating public business information as part of the company’s competitive infrastructure, not just as website copy or marketing material.

A small business should be able to answer basic questions clearly across its website, public profiles, listings, and customer-facing channels:

The next step is to strengthen the signals that support those answers. Businesses can update product and service pages, add clear pricing or policy information where appropriate, collect reviews across relevant platforms, keep business listings consistent, publish useful content that demonstrates expertise, and look for credible third-party mentions from industry sites, local organizations, partners, media outlets, or professional directories.

For retail companies, that may mean clearer product descriptions, better product data, stronger review collection, transparent shipping and return policies, and more consistent listings across marketplaces and shopping platforms.

For service businesses, it may mean clearer service-area pages, case studies, testimonials, staff expertise, licensing information, appointment details, response-time expectations, and customer reviews on the platforms where buyers already search.

Direct customer relationships also become more important. Retailers can build email lists, loyalty programs, repeat-buyer campaigns, private communities, and referral programs. Service businesses can do the same through newsletters, customer follow-ups, maintenance reminders, educational content, local partnerships, and referral networks.

The companies that depend only on traditional search traffic may be more exposed as AI-driven discovery grows.

That may be one of the most important lessons from media. If traffic from search engines and AI systems becomes less predictable, owned relationships become more valuable.

The future of AI-driven discovery will depend on several unresolved questions:

Prince also raised a related advertising question: what happens when the visitor is not a human, but a bot or AI agent? Traditional display ads were designed to influence people. Agentic commerce may require a different system for informing, influencing, or qualifying AI agents, especially if paid ad placement becomes part of AI-generated recommendations.

AI-driven discovery is still very early. No company can know exactly how agentic search and commerce will develop. For now, the practical work is to improve website structure, credibility signals, citations, reviews, and public information so a business has a better chance of being found, evaluated, and recommended.

Prince’s larger point is that the signals businesses use to build trust may need to change in an agent-driven market.

For human customers, trust can come from emotion, design, familiarity, loyalty, advertising, and the feeling a company creates. For AI agents, trust may become more tied to evidence: reviews, fulfillment history, policies, citations, business data, availability, customer experience, and other signals that can be measured or compared.

For businesses, that would change how trust is built. The companies that understand which signals influence AI recommendations may have an advantage over those that still treat brand visibility as something built only for human attention.

Q: What did Matthew Prince say about AI and small businesses?
A: Cloudflare CEO Matthew Prince warned that AI could “destroy small businesses” if AI agents become a major way consumers discover, compare, and buy products or services. His concern is that agents may favor established companies with more public trust signals, making it harder for smaller or newer businesses to compete.

Q: What is agentic commerce?
A: Agentic commerce is the use of AI agents to research, compare, recommend, or complete purchases for consumers or businesses. Instead of a person browsing many websites, an AI agent may narrow the options before the customer sees them.

Q: Why could AI agents make small businesses harder to find?
A: AI agents may rely on evidence such as reviews, citations, pricing, availability, delivery history, return policies, customer service records, and structured business information. Larger companies often have more of those signals online, which could make them easier for AI systems to evaluate and recommend.

Q: Why are publishers seeing this problem first?
A: Publishers are seeing the issue first because AI answer engines can summarize articles and reduce the need for users to click through to the original source. That makes media an early example of how AI can use information from a website while sending back fewer visitors.

Q: Is AI discovery just another version of SEO?
A: AI discovery overlaps with SEO, but it is not the same. SEO helps businesses compete for ranking in search results. AI agents may recommend only a few options, summarize the market, or complete a purchase directly, which makes selection more important than simple ranking.

Q: Could AI help small businesses instead of hurting them?
A: Yes. AI can help small businesses reduce costs and increase capacity in areas such as marketing, writing, design, coding, research, customer service, administration, and data analysis. The challenge is that becoming more efficient does not automatically make a business easier for customers or AI systems to find.

Q: Are local businesses safer than online businesses?
A: Some local and experiential businesses may be more resilient because they rely on proximity, personal relationships, service quality, and community reputation. But local businesses may still be affected if customers use AI agents to find restaurants, repair shops, agencies, doctors, attorneys, accountants, contractors, or other service providers.

Q: What should small businesses do now?
A: Small businesses should strengthen the public signals that help customers and AI systems understand, verify, and trust them. That includes clear websites, structured information, consistent business listings, strong reviews, third-party citations, transparent policies, useful content, and direct customer relationships.

Cloudflare’s CEO Matthew Prince’s warning about AI and small businesses comes down to customer access. As AI agents influence how people search, compare options, and make purchasing decisions, the businesses that are easiest for AI systems to verify may be the ones customers see.

AI may make small businesses more productive while making discovery more demanding. The same tools that help small teams write, research, design, code, market, and serve customers may also create a market where visibility depends on reviews, citations, structured data, clear policies, and other public proof of trust.

Small businesses, publishers, retailers, service providers, ecommerce platforms, payment networks, marketplaces, and AI companies all have a stake in how this develops. Any business that depends on search, referrals, reviews, local listings, or online reputation could be affected if AI systems begin deciding which companies customers see, compare, or consider.

This matters for competition because customer acquisition may move from being found in search results to being selected by AI systems before the customer sees the full market. If those systems favor companies with the most existing data, reviews, citations, and fulfillment history, larger established companies may become easier to recommend by default.

For small businesses, the main decision is whether to wait and see how AI-driven discovery develops or start building the public trust signals AI systems may use to find, compare, and recommend them.

In short, AI may help small businesses do more with less, but it may also raise the standard for being found, trusted, and chosen through clearer public trust signals.

Businesses that prepare now may be easier for AI systems to recommend later. Businesses that wait may learn too late that customers are no longer finding them.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing support, AEO/GEO/SEO optimization, image concept development, and editorial structuring support from ChatGPT, an AI assistant. All final editorial decisions, perspectives, and publishing choices were made by Alicia Shapiro.