Articles

On Monday morning, a sales rep named Sam needs to reach the owner of a small heating and cooling company. Sam, the owner, and the company are all made up. So Sam asks AI for a friendly, personalized cold email to Dana, who runs Northline HVAC. Ten seconds later, he has this: Hi Dana, I came across Northline HVAC and was impressed by your great reviews and strong reputation. We help businesses like yours grow with innovative, AI-powered solutions. Would you be open to a quick call this week? Polite. Friendly. Dana’s name in the first line. And nothing else. Sam hits send. By Friday, Dana still hasn’t answered. Nobody did anything wrong. That’s what makes it so hard to fix. At BetterMarketers, we write for people who sell for a living, so this is the version of the problem we care about most: the message that sounds right and goes nowhere. AI can now produce that message by the thousands.

A man walks into a hardware store and asks for a quarter-inch drill bit. The salesperson shows him four. Sizes, materials, prices, cases. Then a second salesperson asks, “What are you making?” “A shelf. My wife has been asking since March.” Now the second salesperson is selling something different: a shelf, up on the wall, before the weekend. (This is a well-known marketing story in different versions. We’re telling it our way.) He never wanted the bit. He wanted the shelf, and an end to the nagging.

On Friday afternoon, Carlos presents the monthly report. He’s made up. Traffic is up 40%. Social followers are up 25%. Leads are up 75%. Slide after slide of green arrows. The CFO asks one question: “So why is revenue flat?” Carlos doesn’t have a slide for that.

Lena needs a gift for her father, who likes to cook but hates clutter. She’s made up. She doesn’t open Google. She asks an AI assistant: “Gift under $80 for a dad who cooks, doesn’t want gadgets.” It gives her three options with short reasons. She taps one, checks the price, and buys. The brand never got a pageview. It also never saw her scroll past its competitors.

Two emails land in Ben’s inbox on the same morning. He’s made up. Email one: “Hi Ben, we’ve got something special for you!” Email two, from a company whose running shoes he looked at twice last week: “Still thinking about the trail shoes? Here’s how they fit if you have wider feet, from people who bought them.” Both used his name or his data. Only one felt like it was written by someone who noticed him.

Priya is shopping for a project management tool for her six-person agency. She’s made up. She visits three vendor sites. All three say “streamline your workflow.” So she does what a lot of buyers do. She opens a forum and types: “Is [tool] worth it for a small agency? Honest answers only.” Forty minutes later she’s read two complaints about pricing, one story about slow support, and one reply from a stranger who says, “We switched last year. Here’s what I wish I’d known.” She buys from the vendor with the stranger’s good review. She never filled out a form.

Nia has a content calendar color-coded like a rainbow. (She’s made up. She runs social media for a software company that makes scheduling tools for clinics.) Forty posts a month. Clean graphics. A strong voice. Last quarter, one post hit 80,000 views. Her boss asks one question: “How many demos came from that?” Nia checks. The answer is two. One was her cousin. Views are not the problem. Nia’s posts are being watched by the wrong people, for the wrong reason, with nothing to do next.

Two sales teams start Monday. At team A, a rep opens her laptop and finds a short brief for each of today’s calls: what the company does, what changed recently, who else is on the buying committee, and a draft follow-up waiting for her review. At team B, a rep opens six browser tabs. LinkedIn, the company website, the CRM, last week’s email, a spreadsheet, and a note that says “call Dave.” Both reps are good. One is ninety minutes ahead before the first call. (Both teams are made up. The ninety minutes is an illustration, not a statistic.)

Imagine Jonas. He’s not real. Jonas owns a plumbing company. For years, “water heater repair near me” brought him about forty visitors a month, and a few of them called. This year the visitors dropped. His rankings didn’t. He checks Google himself. Above his listing, a summary answers the question: signs your water heater is failing, what repairs cost, when to call a pro. The searcher got what she needed. She never clicked. Did Jonas lose a customer? Or did he lose a visit? Those aren’t the same thing. How many clicks are disappearing? Less than the scariest headlines say, but more than anyone wanted. Pew Research Center tracked the browsing of 900 U.S. adults in March 2025. On Google pages with an AI summary, users clicked a traditional result link in 8% of visits. On pages without one, it was 15%. That’s roughly half as often. Only 1% of visits clicked a link inside the summary itself. (Pew says this is an association, not proof the summary caused it.) Ahrefs looked at 300,000 keywords and reported that when an AI Overview shows up, the top result’s click-through rate was 58% lower than what it would have been otherwise, comparing December 2023 to December 2025. That “would have been” is a model, and the data is desktop only. Ahrefs itself says it’s a correlation, and other studies landed between roughly 47% and 65%. So the honest summary: clicks are falling, on certain queries, by a lot. The exact number depends on who measured and how. Is a lost click a lost customer? Sometimes. Often not. Think about who clicks on “what is a water heater flush.” Someone doing homework. Not someone about to hire. Now think about “emergency plumber open now.” Someone ready to pay. That person still clicks or calls. Here is an opinion, not a finding: the clicks that disappear first are the easiest ones, the questions a summary can answer. The clicks that remain are more likely to come from someone who wants a person. If that holds for you, you lose traffic and keep buyers. Check your own numbers before you believe it. Compare leads, not visits. What should Jonas do? Be the name in the answer. If the summary mentions businesses, you want to be one. That comes from the same things that build trust anywhere: a specific, honest website, strong reviews, and being talked about by others. Make the click worth it. If someone does click, don’t greet them with a blog post. Give them the next step: a phone number at the top, a price range, a booking button. Write what a summary can’t. A summary can explain a flush. It can’t show your actual before-and-after photos, your actual prices, or your technician’s name. Put those on the page. Ask how they found you. Add one field to every form: “How did you hear about us?” Searchers who never clicked still show up eventually, as a call or a search for your name. Watch branded searches and direct traffic. If people see your name in an answer and later type it, that shows up there, not under organic. The part nobody likes to say Some of this is simply loss. If your business made money from answering simple questions with ads on the page, a summary took that. Don’t pretend otherwise. But if you sell something a person has to choose, trust, and buy, the job didn’t vanish. It moved earlier. The first impression now happens before your site loads. BetterMarketers readers in sales feel this fastest. Your prospect often arrives already half-decided. To see how Google’s AI is changing the funnel, read our guide “Google’s AI Just Killed the Marketing Funnel — What to Do Next”: https://bettermarketers-web.vercel.app/g/1GgWTZuk What Jonas should count Jonas doesn’t need forty visits back. He needs to know how many calls came from the ones who did click, and whether the name Jonas Plumbing shows up where people are asking. Count those two things for a month. Better Said AI summaries mean fewer clicks, and the research shows it. But lost clicks aren’t always lost customers. Track leads instead of visits, make your remaining clicks easy to act on, and build a name that shows up in the answers.

Marcus has a goal for the quarter: ten meetings. So on Monday he does what the tools tell him to do. He uploads a list of 1,000 owners of independent dental practices and tells the AI to write a “personalized” email for each. It finishes before his coffee does. Here’s one of them: Hi Dr. Patel, I saw that your practice has been serving patients for years and thought you might be interested in how our platform helps practices like yours streamline operations. Do you have 15 minutes this week? Marcus hits send on all 1,000. Friday comes. Eleven replies. Nine say “unsubscribe.” (Marcus, Dr. Patel, and the numbers are made up. The feeling is not.)

Meet Rosa. She’s not real, but you know her. Rosa runs a small bookkeeping firm for dentists. Her website sits near the top of Google for “dental bookkeeping.” She’s proud of it. She paid for it. One evening, a dentist named Tom opens an AI assistant and types: “Who should I hire to do bookkeeping for my practice?” He gets three firm names and a short reason for each. Rosa isn’t one of them. Tom never searched Google. Rosa never knew there was a race. Is this really happening? Search interest in it is, at least. Ahrefs reported in July 2026 that searches for “generative engine optimization” grew 997% over the last 18 months. (The post gives growth, not a monthly volume, so we can’t tell you how big the base was. Small numbers grow fast.) Generative engine optimization, or GEO, just means trying to show up inside AI-written answers. It’s the same job as SEO, pointed at a different door. Here’s the part worth sitting with. Ahrefs’ own post questions what AI citations are worth. One SEO, Wil Reynolds, reported a 1,900% jump in ChatGPT citations to a page on his site and saw little to no business impact. So visibility alone isn’t the prize. Being mentioned for the right question, by a buyer who is ready, is. Why would an AI pick one business over another? Nobody outside these companies knows the full recipe. Anyone who says they do is selling something. What we can say, from public research: AI answers lean on what’s already written about you across the web, not just on your own site. Our read, and it’s ours, not Ahrefs’: an AI system looking for “who should I hire” wants more than one voice agreeing. Your own homepage saying you’re great is one voice. What can Rosa do? None of this guarantees a mention. It just makes a mention more likely than doing nothing. Say exactly who you serve. “Bookkeeping for small business” fits ten thousand firms. “Bookkeeping for independent dental practices, including insurance reimbursement tracking” fits a few. Specific pages answer specific questions. Make your facts match everywhere. Same name, same services, same city, same description on your site, your LinkedIn page, your Google profile, and your directory listings. If three places say three things, a machine has to guess. Get written about by other people. Reviews with details. A guest article on a dental association’s blog. A client who mentions you in a forum thread. You can’t force this. You can make it easy and earn it. Check what AI says about you. Ask the same buyer question in a few assistants. Write down who gets named and why. Do it monthly. It takes fifteen minutes. Nobody has a proven playbook that works every time. Another study cited inside that Ahrefs post found a page cited 13.3% of the time before heavy “GEO” changes, and 10.9% to 12.2% after. More effort, slightly lower number. One study, one model, so don’t over-read it. But it’s a good reminder to test, not assume. What does this change for salespeople? At BetterMarketers, we write for people who sell for a living, so here’s the sales angle: Your buyer may arrive already holding a shortlist an AI gave them. If you’re on it, you start warm. If you’re not, you can’t see what happened. Ask new prospects one question: “How did you start looking for someone like us?” The answers will tell you more than your analytics will. To see what Google’s AI changes about how buyers find you, read our guide “Google’s AI Just Killed the Marketing Funnel — What to Do Next”: https://bettermarketers-web.vercel.app/g/1GgWTZuk Back to Rosa Rosa doesn’t need a new website. She needs three dentists to say, in public and in their own words, what she fixed for them. Then she needs to check, every month, whether the machines noticed. Better Said Ranking on Google and being named by an AI are two different results. Use specific pages, consistent facts, and real third-party mentions to improve your chances, then check AI answers monthly. No tactic guarantees a recommendation.