Category: Executive Briefings

When AI answers the question directly, there are no click-throughs. Most owners don’t see the damage until it shows up in a down month – and by then, it’s too late.

• Roughly 60% of searches now end without a click to any website. When AI Overviews appear, click-through rates on organic results fall by about 58%.
• Standard analytics platforms do not measure the buyers who never arrived. The damage often surfaces as soft demand for two to three quarters before it is correctly identified.
• The same shift is producing asymmetric outcomes inside the same categories. Brands cited by AI are gaining qualified traffic that converts roughly 31% higher than legacy channels. Brands not cited are losing both the traffic and the channel.
• The cost runs through three layers – awareness, customer acquisition cost, and brand asset value. Each layer compounds and makes the next more expensive to fix.
• The first move is diagnostic, not promotional. Audit the questions your buyers actually ask AI – and whether you appear in the answers.

The shift is happening quietly

Last quarter, a CEO of a 22-year-old specialty ecommerce business called us. Traffic was down 22% year over year. Revenue was off 4%. The brand was operating exactly as it had the year before – same product mix, same fulfillment, same paid spend, same email program. Something was siphoning the top of the funnel and he couldn’t identify the leak.

Nothing was broken. But an entire buyer interaction happened without him knowing it.

Buyers ask AI a question. AI answers it. The answer comes from somewhere – perhaps the client’s site, perhaps a competitor’s, perhaps a manufacturer’s spec sheet or a Reddit thread. The buyer gets what they needed and never visits anyone. The session doesn’t exist. The cart doesn’t exist. The data doesn’t exist.

The industry word for this is zero-click. In Q1 of 2026, roughly 60% of searches ended without a click to any website. When an AI Overview appears on a Google query, click-through rates on the underlying organic results drop by about 58%. Across the entire ecommerce category, ChatGPT referral traffic grew by more than 1,000% in a single year. These numbers describe a structural change, not a cyclical one.

Why your dashboard is the last place this shows up

The most operationally dangerous part of zero-click is that it does not appear in the place most operators look first. Google Analytics counts the buyers who arrive. It cannot count the buyers who never needed to arrive. Search Console reports impressions, which include the queries where you are visible inside an AI Overview – but it also reports clicks, which fall sharply on those same queries.

The CMO sees that impressions are holding steady. The CFO sees that revenue is softening. Marketing concludes something is wrong with conversion. Operations concludes something is wrong with the product. But they’re all looking at the wrong layer.

The AI transition is affecting businesses differently

Zero-click is often described as a uniform headwind. It is not. The same shift is producing wildly different outcomes inside the same categories. Across DTC, industrial, service, and specialty product businesses, three patterns are emerging.

• 1. Brands AI cites are gaining traffic from new channels – ChatGPT, Perplexity, Claude, Gemini – that did not exist 18 months ago. The volume is smaller than legacy Google referral but converts at meaningfully higher rates, often 25 to 35% above the brand’s blended average. AI-referred buyers arrive late in the funnel, ready to act.
• 2. Brands AI bypasses are losing organic traffic to zero-click and not picking up AI referral to replace it. The buyer is getting answered and then either going to a different brand or going direct to a marketplace. Paid media costs rise to compensate. The blended customer acquisition cost climbs even when the marketing program is unchanged.
• 3. Brands stuck in-between appear inside AI answers occasionally, in low-intent queries, but are not consistently recommended for buyer-intent queries. These businesses tend to be doing the right structural work but inconsistently, with a content footprint AI can partly read but not confidently recommend.
• The defining factor is rarely the size of the business or the strength of the brand. It is the structural readability of the content and product data.

The three layers of cost

The cost to your business happens in three layers, each harder to reverse than the last.

Layer one: awareness.

The buyer who would have learned you exist now gets an answer that does not mention you. Even if the category is growing, the number of times new buyers are discovering your brand is decreasing.

Layer two: customer acquisition cost.

When AI does not recommend you, paid media has to buy back the visibility AI was supposed to provide for free. Heavier reliance on paid media increases your overall customer acquisition cost, making it harder to compete with companies that have more significant organic traffic.

Layer three: brand asset value.

The organic equity built over 10 or 15 or 25 years – the rankings, the backlinks, the trust signals – was a competitive moat. It was also an asset on the balance sheet of any future exit conversation. AI is the bridge that crosses that moat.

Each layer makes the next more expensive to fix. A brand that has lost awareness for two years cannot rebuild it with a quarter of paid spend. A brand with elevated customer acquisition costs cannot return to old margins with a single optimization sprint.

How to find out where you stand

The diagnostic is not complicated. It is uncomfortable. Take 10 to 20 questions your real buyers actually ask before they consider buying from you. Not the keywords your old SEO report tracks. The questions a person would actually type into ChatGPT, Gemini, Claude, or Perplexity – comparison queries, problem queries, recommendation queries.

Run them. See what comes back. Note which brands are cited, in what order, with what context. Note where you appear and where you do not. Note where a competitor you did not previously consider a threat is now the named recommendation.

The result is rarely flattering. The brands doing the structural work – clear product descriptions, depth of category context, real first-party data, citation-worthy expertise are the ones that show up. The brands that have been buying impressions and calling it marketing do not.

The down month is the message

There is a temptation, when revenue softens, to look for the cause inside marketing – a tactical fix, a paid campaign, a creative refresh. In the right environment, that approach works. In this environment, it is treating a structural shift as a tactical problem.

The work to be recommended by AI is not new work. It is the work organic search has always rewarded: clarity, depth, trust signals, a structure that machines can read and humans can act on. What changed is the audience reading it. The operators who hear that message early have time. The ones who hear it won’t.

“An off month isn’t the problem. An off month is the message.”

AI is starting to make recommendations and complete purchases for your customers. That has real consequences for your margin. Let’s look at what’s changing and how to prepare.

• AI shopping isn’t hypothetical anymore. Amazon, ChatGPT, Walmart, and Shopify have all shipped AI-driven purchase experiences in 2026.
• When AI mediates the purchase, three things change about your DTC unit economics: discovery moves upstream of your website, comparison happens on what AI can read, and AI decides which competitors you get compared to.
• The assets you’ve used to defend premium pricing – brand storytelling, premium photography, conversion-optimized landing pages – don’t survive an AI summary. They show up after the recommendation, by which point the comparison is over.
• Three things protect your margin: detailed product data, brand context AI can carry, and first-party customer relationships AI can’t mediate.
• Most DTC operators have 12 to 24 months before AI mediation is dominant in their category. The work to prepare is straightforward. The window won’t stay open forever.

The shopping moment is moving off your site

For most of the history of DTC, the contested moment was your website. The buyer landed, your page told the story, you earned the conversion. Most of what DTC marketing has been about is optimizing that moment – the hero image, the headline, the social proof, the path to cart.

That moment is moving in 2026. Amazon’s Rufus is recommending products inside the Amazon experience, before any product page click. ChatGPT is completing transactions inside the chat. Walmart has shipped conversational shopping. Shopify has rebuilt its developer platform for agentic checkout. Klarna is positioning as a shopping agent. Perplexity is testing buy-from-search.

None of this is speculative. These are shipped products at the platforms that handle most U.S. commerce volume. What’s uneven is how many buyers in your specific category are routing their purchase through an AI layer instead of coming direct to your site. In some categories, the number is already meaningful. ChatGPT-driven ecommerce referral traffic grew more than 1,000% in 2025.

The question isn’t whether AI shopping is coming to your category. It’s whether your margin survives when it does.

Three things that change about your unit economics

When AI mediates the purchase, three things change about the economics of a DTC sale. Each matters on its own. Together they reshape what your operating model has to do to protect your margins.

1. Discovery moves upstream of your website.

Your first impression used to be your home page. When AI makes the recommendation, your first impression is the sentence AI produces. If that sentence contains your brand with a confident description, you’ve earned the consideration moment. If it doesn’t – or if AI describes you inaccurately or alongside three commodity competitors – your website has to work much harder to recover.

Right now, AI is describing most premium DTC brands in ways that range from imperfect to actively unflattering. The premium isn’t in the description. The attributes are often slightly wrong. The positioning groups them with competitors they wouldn’t call alternatives. This is the new top of your funnel, and it’s happening before any analytics can see it.

2. Comparison happens on what AI can actually read.

Your premium was built on assets AI doesn’t read well. Brand storytelling collapses into descriptive text. Premium photography reads as the existence of an image, not as a quality signal. Brand voice gets summarized, not experienced. The atmospherics that justified a 2x price premium don’t survive an AI summary.

What AI reads well: product specifications, materials, dimensions, use cases, comparison tables, structured reviews, third-party citations. The parts of a product description that look ordinary on your website become disproportionately important when they’re the inputs into a recommendation. The brands winning in AI-mediated commerce are doing both – story and structure – and publishing both in formats AI can read.

3. AI decides who you get compared to.

AI categorizes your product and picks reasonable alternatives. A brand that spent a decade differentiating itself from a set of competitors can find itself recommended alongside three of them, because AI grouped them as alternatives. Your premium has to be defensible at the level of the AI recommendation, not just at the level of your website. Brands that publish their own perspective on what category they’re in and what makes them different have meaningful protection. Brands that let AI categorize them get commoditized in the recommendation.

What protects your margin

Three things defend DTC margin when AI is mediating the purchase. None is new. All are more important than they were 18 months ago.

Detailed product data.

Specifications. Materials. Dimensions. Use cases. Ingredients. Comparisons. Written for both buyers and the engines that read them. The brands that give AI more to work with get described more accurately and recommended more often. This is unglamorous work. It’s also the highest-leverage work most DTC brands can do this year. The standard for product data in a pre-AI world was sufficient. The standard now is comprehensive.

Brand context AI can carry.

The reason your brand exists. Your point of view. The story that justifies your premium. Published alongside your product data, not buried on a separate about page that AI won’t cite when it’s making a recommendation. AI can carry brand context if you give it enough material. It can’t infer it. This is where most premium DTC brands are getting hurt right now – the brand story exists, but it’s structurally invisible to AI because AI can’t find it in the content it actually reads.

First-party data and direct customer relationships.

Email subscribers. SMS. Loyalty. Owned community. Repeat purchases. Customer service. The parts of your DTC relationship AI doesn’t get to mediate. The point now is that the strategic value has gone up again, for a different reason. AI shopping makes brand-to-buyer mediation harder. Your first-party relationships are the places you’re still talking to your customer directly, without an intermediary deciding what they hear.

“AI shopping is coming. The only question is whether your margin survives it.”

How you structure your support resources is critical to your success in a rapidly changing marketplace. Let’s look at it from an operator’s perspective.

• Deciding what to build in-house, what to buy off the shelf, and what to bring in a partner for is one of the harder calls you’ll make as a CEO.
• Three questions, in order, can help you work through it: is this capability core to who we are, how fast is it changing, and what does it cost to get wrong?
• Most cost comparisons get the math wrong. The salary isn’t the full cost of building. The subscription isn’t the full cost of buying. A good partner is often cheaper than either when you load it all up.
• Most healthy businesses use all three. Build the core. Buy the routine. Partner the disciplines that are deep and changing fast.
• For now, AI visibility belongs in the partner bucket for most founder-led businesses. The field is moving too fast to keep up from inside one company.

The question every operator is asking now

For founder-led businesses today, successfully navigating AI’s transformation of search and discovery platforms is critical to their long-term success. The question is “What’s the right mix of internal and external resources to keep me ahead of my competition?” 

Why this decision usually gets hijacked

Most of the time, this comes down to a few conversations with people you trust.

Your head of marketing wants to hire someone. Owning the capability matters, they’ll say, and they’re right. They also have a job that depends on the capability being inside.

Your CFO likes the look of a software subscription. Lower headcount, predictable cost. Also right. Also a tidy line item.

A consultant you’ve worked with pitches taking it on themselves. Depth and speed neither a hire nor a tool can match. Also right. Also their revenue.

None of them is lying. Each is genuinely arguing for what they believe – and what they happen to benefit from. Your job is to weigh all three honestly and pick what’s best for the business.

Three questions, in order

Three questions can help you work through this. The order matters – answer them out of sequence and it’s easy to talk yourself into the wrong answer.

1. Is this a core competency?

A core competency is the thing that makes your business yours. Your brand. Your customer relationships. Your product. The way you sell. If competitors can’t easily copy it, it’s core.

If it’s core, you build it. You don’t outsource what makes you who you are.

2. How fast is the capability changing?

This is where most build decisions go wrong. Hiring works when the field is stable – when the person you hire today is still at the leading edge in 18 months. Accounting is stable. Fulfillment is stable. Parts of product development are stable.

AI Is not. In the last six months alone, buyer behavior has changed for hundreds of millions of consumers. More than 50% of all Google searches today end with zero clicks. Protocol standards, natural language search tagging, and structured data changes have to be made on a weekly and monthly basis. Keeping up with this pace of change is a team effort, not something that individual employees can possibly manage on their own.

3. What’s the cost of getting it wrong?

The cost to your business happens in three layers, each harder to reverse than the last.

Layer one: awareness.

The buyer who would have learned you exist now gets an answer that does not mention you. Even if the category is growing, the number of times new buyers are discovering your brand is decreasing.

Layer two: customer acquisition cost.

When AI does not recommend you, paid media has to buy back the visibility AI was supposed to provide for free. Heavier reliance on paid media increases your overall customer acquisition cost, making it harder to compete with companies that have more significant organic traffic.

Layer three: brand asset value.

The organic equity built over 10 or 15 or 25 years – the rankings, the backlinks, the trust signals – was a competitive moat. It was also an asset on the balance sheet of any future exit conversation. AI is the bridge that crosses that moat.

Each layer makes the next more expensive to fix. A brand that has lost awareness for two years cannot rebuild it with a quarter of paid spend. A brand with elevated customer acquisition costs cannot return to old margins with a single optimization sprint.

Measure the true costs

When you sit down to run the numbers, the spreadsheet usually misses the real costs. Not on purpose – the easy-to-see numbers just aren’t the full picture.

Build, fully loaded.

The salary is just the start. Add benefits, payroll tax, recruiting (often 20-25% of first-year comp), three to six months of ramp time, management overhead, tools, turnover when the hire eventually leaves (18 to 36 months in fast-moving digital work), and the ceiling that comes from one person owning a discipline that requires exposure across many businesses to stay current. Fully loaded, build runs 2x to 2.5x the headline salary.

Buy, fully loaded.

The subscription is just the start. Add implementation (often four to twelve weeks), integration with your systems, the internal time to interpret what the tool produces, and the eventual cost of switching tools later. Tools report and alert. They don’t think for you. They don’t know your business. The thinking still has to happen somewhere.

Partner, fully loaded.

The retainer is just the start. Add your internal time to brief and review, relationship overhead, and the risk if the partner turns out to be the wrong one. Against fully loaded build and fully loaded buy, the partner option is often the cheapest of the three for the value it delivers – especially in fast-moving disciplines. The spreadsheet doesn’t show this because it rarely loads the alternatives correctly.

The hybrid most healthy businesses run

Most healthy businesses we see use all three. The real question isn’t “build, buy, or partner” – it’s which capabilities go in which bucket. Here’s the way we see it:

• Build the things that are structurally yours – brand voice, customer relationships, the disciplines that compound over years of consistent investment.
• Buy the tools that handle routine, well-defined work – analytics, email, CRM, helpdesks. Standardization is the point.
• Partner the disciplines that are deep, fast-moving, or hard to staff – AI visibility, paid media operations, technical specialty work.

Where AI visibility lives right now

If you’re trying to figure out where AI visibility falls right now, the honest answer for most founder-led businesses is partner. Not forever. But for now.

The field is moving too fast for one person inside one business to keep up. Good practitioners leave single-business roles within 18 to 24 months because the work narrows too quickly without exposure to other categories. The tools are useful but require judgment they can’t provide. And the cost of being structurally less visible to AI than your competitors compounds in a way that’s hard to reverse. Finding the right partner at this stage is the key to leveraged growth.

“Build the core. Buy the routine. Partner the deep and fast-moving work.”

The pattern is consistent: the work moves to a junior or outsourced team, gets gradually thinner over time, and ends in disengagement. Here is how to recognize it before you sign – and find a better partner for your business.

• In any room of 50 CEOs, roughly 45 will raise a hand when asked if they have had a bad experience with a digital marketing agency. Unfortunately, this is just he baseline condition of the industry.
• The pattern is consistent across categories: the work moves to a junior or outsourced team carrying too many accounts, then gets gradually thinner over time. It repeats because the agency model is structurally designed to produce it.
• Average client tenure in this category is roughly six to nine months – too short to compound returns, too long to easily walk away.
• The structural fix is a different operating model: senior-led delivery, small account teams, transparent staffing, no service relabeling. That model exists. It’s rare because it’s harder to scale.
• The questions a CEO should ask before signing are not about capabilities. They are about who will actually be doing the work, how performance will be judged and what outcomes to expect.

The 90% is real

Stand at the front of a room of 50 founder-operators and ask them to raise a hand if they have had a bad experience with a digital marketing agency. Nearly every hand goes up. We’ve asked this question to a lot of CEO forums – it’s roughly 90% every time.

That number is not sample bias. It tracks with published benchmarks on agency client retention, satisfaction scores, and agency-to-agency churn in the small and mid-market segment. The pattern is consistent across categories too – ecommerce, industrial, service, brand-led DTC, multi-business portfolios. It’s consistent because it is structural.

The five-part pattern

To better understand how this happens we can look at the process as five distinct parts. Most CEOs recognize the first when they are inside it, the second in retrospect, and can’t prevent the third, fourth, and fifth without leaving the engagement.

Part one: the sale.

The pitch is led by polished, experienced people. The deck is dialed in, the case studies are real, the questions are smart. The contract is signed on the credibility of the people who showed up.

Part two: the handoff.

The account moves to a delivery layer the client rarely sees described accurately in any capabilities deck. In most small and mid-market digital agencies, that delivery layer is some combination of offshore writers, designers, and analysts in the Philippines, India, or Eastern Europe; freelance contractors brought on per-project through Upwork, Fiverr, or specialty marketplaces; and junior in-house staff 18 to 36 months into their careers.

None of this staffing structure is disclosed in advance. There is no slide in the capabilities deck that names the freelance platform. There is no line in the contract that says 60% of execution will be outsourced offshore. The first 60 to 90 days consist of onboarding deliverables: audits, kickoff decks, strategy documents, planning workshops – rarely the work the CEO thought they were buying.

Part three: the relabeling.

As the market shifts, agencies relabel existing services as the new category. SEO became content marketing. Content marketing became inbound. Inbound became growth marketing. Growth marketing became digital transformation. Today, the wrapper is AI. In many cases the underlying work has not changed. Same sub-par work, just with a different label.

Part four: cost hike.

Results are softer than the pitch implied. The agency recommends an additional service: paid social, conversion optimization, brand strategy. They tell you you need to spend more on ads to keep up. Total monthly cost climbs 40 to 60% above the original contract. The results are still soft. The conversation about results has been replaced by a conversation about activity.

Part five: the gradual reduction.

Agency costs climb every year. Wages, tools, leases, insurance. Revenue from any single client does not move at the same rate. The agency has one structural lever to maintain margin against rising costs: do less work for the same price.

The reduction is gradual. A monthly report that used to be custom-built becomes a templated deliverable with the client’s data dropped into a standard format. A quarterly strategy session becomes a meeting where last quarter’s deck is updated with new numbers. An audit that used to take 30 hours becomes a 6-hour pass through a standard framework. Custom content becomes a lightly edited template, increasingly drafted by AI. A 90-minute call gets rescheduled to 30. Senior involvement that used to happen weekly happens monthly, then quarterly. Reports become more generic. The strategist who used to attend monthly reviews is replaced by an account coordinator.

The retainer stays the same. The labor input shrinks. The margin is preserved. The CEO realizes the agency they signed and the agency they have are no longer the same, but can’t point to the moment when it changed. Most describe it as “the agency just isn’t bringing the energy anymore.” It is rarely an energy problem. It is a labor problem.

Why the model produces this

The pattern persists because the underlying economics push toward it. Agency margin comes from the gap between what senior people are billed at and what junior or contracted people are paid to do the work. The healthier the margin, the more leveraged the delivery. Most contracts are priced on hours or capacity, not outcomes – so the economic incentive is to consume hours, not produce results. And as services commoditize, the path to maintaining margin is relabeling: calling the same work by a more expensive name while the actual labor input declines quietly underneath.

None of this is malice by individuals inside the agency. It is the structural response of a business that cannot raise prices on existing contracts and cannot reduce its cost base. The only variable available is the labor input on each account. The agencies that are most disciplined about preserving margin are also the ones most disciplined about shrinking that input over time.

The structural alternative

A different operating model exists, with two easily recognizable characteristics:

• Named, transparent staffing. You knows who is doing the work, what their level is, and how the labor is allocated. The account manager you meet in the sales process is the same person you’re talking to 6 months in.
• Monthly discussions are focused on measurable business outcomes the client can verify – revenue, leads, citation share, margin – not hours or activity.

The questions that separate the two models

Every agency can present a capabilities deck that looks similar. Here are the questions that matter:

• 1. Who specifically will be doing the work on my account? Names, titles, years of experience. What percentage is in-house versus outsourced or contracted? Not a team chart. Specific humans.
• 2.  What outcomes will we be measuring? How will they be verified?
• 3. What will you tell me when we are losing? A partner who has only delivered wins is either lying or new. Real partners have had hard quarters and learned from them.

The answers tell you whether you are buying a relationship or just another pitch.

“The 90% number is not a problem the industry will fix on its own. The industry is the problem.”

It is not about rankings anymore. AI recommends the brands whose products are structured so it can read, understand, and match them to what a buyer actually asked for.

• AI recommendation is not algorithmic ranking. It is confidence-weighted synthesis. The brands AI cites are the brands AI can describe accurately.
• The classic SEO triangle – relevance, authority, technical – is being replaced by a different triangle: machine-readable structure, first-party authority, and brand context AI can carry.
• Some 20-year market leaders are disappearing from AI answers while smaller competitors with cleaner content footprints are taking their place. This is happening across categories.
• The strongest predictor of AI citation is not budget or backlinks. It is the depth and clarity of category context the brand publishes about itself.
• For brand builders, the risk is brand erosion through over-optimization. The opportunity is that AI can carry brand context – if the brand provides the right structure.

Twenty years ago you ranked. Then you got recommended.

This is not a phrasing change. It is the shift underneath the entire current conversation about AI search.

For two decades, getting found online meant ranking. Search engines were ranking machines. They asked a single question in many variations: which page on the web is the most relevant answer to this query, weighted by signals of authority? The winners published pages with the right keywords, secured backlinks from credible domains, and maintained the technical foundation that signaled a real business behind the URL.

AI does not match a page to a query. It matches an answer to a question. And the answer is assembled from many sources at once. That is a structurally different problem, and it produces structurally different winners.

The recommendation stack

When a buyer asks ChatGPT, Gemini, Claude, or Perplexity for a recommendation, the model is not picking the page with those words in the right order. It is running a sequence of internal operations: understanding the question and its constraints, identifying candidate brands it has substantive information about, filtering by how confidently it can describe each candidate, matching to the buyer’s specific constraints, and producing a short, confident answer naming one to five brands with brief justifications.

The brand that wins the recommendation is not the brand with the highest market share. It is the brand the model can describe with the highest confidence inside the buyer’s specific constraints. That distinction matters enormously. Market share, brand history, and even budget are not the controlling variables. The controlling variable is whether the brand has provided AI engines with enough credible, structured, accurate content to support a confident recommendation.

Why some incumbents are disappearing

Across categories, we are seeing an unexpected pattern. Some 20-year market leaders – brands that dominate Google rankings, brands every operator in the category would name – are showing up rarely or not at all in AI answers. Meanwhile, smaller competitors with much less brand recognition are appearing consistently.

The pattern is not random. The disappearing incumbents share a profile:

• Their content was built for SEO, not for category context. Pages target keywords, not buyer questions.
• Their product data is thin or generic. Specs without context. Categories without rationale.
• Their brand story lives in marketing collateral, not on indexed pages. A buyer can find what they sell, but not who they are.
• Their authority signals are concentrated in backlink count rather than in published expertise.

The smaller brands showing up share an opposite profile. They publish category context – what’s the difference between X and Y, when do you use Z, what should you ask before buying. They use buyer language. They have a clear point of view that distinguishes them. They give AI engines enough first-party material that the model can describe them confidently.

Confidence is the new authority

The classic SEO authority signal – backlinks, domain age, branded search volume – was a proxy for trust. AI uses many of the same signals, but it weights them differently. The new dominant signal is internal confidence: how certain the model is about what to say about the brand.

This is a critical distinction. The model is optimizing to be useful and accurate. If it can describe Brand A with 90% confidence and Brand B with 60% confidence, it will recommend Brand A – even if Brand B is the market leader. The model has no incentive to recommend a brand it cannot describe well. The cost of being wrong is higher than the cost of being incomplete.

The implication is direct: AI is not picking a winner. It is picking the brand it can speak about with confidence. The work is to make that brand yours.

The brand voice trap

For operators whose brand is part of the product, the most important risk is not lack of visibility. It is brand erosion through over-optimization. There is a temptation, once an operator understands how AI engines parse product data, to optimize the brand down to its attribute stack. Materials, sizes, prices, use cases, comparisons. The result is a brand AI can describe confidently – but only as a list of features.

When a buyer asks for a recommendation in a category where the brand was the reason to buy, that strategy fails. AI recommends the brand alongside three commodity competitors with similar attributes. The brand premium evaporates. The story is gone.

The work is to preserve brand context – the point of view, the founder story, the reason the brand exists – in the same content surface where the product data lives. It’s about structuring the data to convey product information alongside brand value.

What practitioners should be doing now

The work of being recommended by AI is the work of being describable. In priority order:

• 1. Audit how AI engines describe you today. Ask ChatGPT, Gemini, Claude, and Perplexity to describe your brand, recommend your category, and compare you to competitors. The answers are the starting point.
• 2. Address the gaps in category context. What questions do real buyers ask before they choose your category? Are those questions answered, in depth, on pages AI engines can find?
• 3. Tighten product and service data. Categories, attributes, comparisons, use cases – structured so machines can read it and humans can act on it.
• 4. Publish first-party authority. Original data, original perspective, original analysis. The material AI engines cannot get from anyone else.
• 5. Protect brand context. The story, the point of view, the reason the brand exists – alongside the product data, not in a separate corner of the site.
• 6. Monitor citations, not rankings. Rankings are a 2010s metric. Citations – who AI names when it answers – are the 2026 metric.

The operators who started in 2024 are already pulling ahead. The ones starting now still have time. The ones who wait until 2027 will be rebuilding from a position significantly behind their newer competitors.

“Your brand story can’t help you if it’s not visible and understandable to AI platforms.”