Digital Desk
Get Started

AI Is Changing How Customers Find You. Here’s What It Actually Means for Your Business.

A plain-English guide to SEO, AEO, and GEO, and what to do about all three.
SEO AEO GEO Guide

By Barbara L. Rogers, Digital Desk Inc.

Originally published September 2026. Last updated October 29, 2026. This space changes quickly, so we revise this guide rather than let it go stale.

If you’ve been hearing that “SEO is dead,” that you need to buy something called GEO, or that AI is about to erase your website traffic, you’re not alone. It’s the most common question we’re getting from clients right now.

Here’s the short version: something real did change; it didn’t change equally for every business. Those businesses that adapt calmly will come out ahead of those that panic.

This guide explains what happened, what it means for your business, and what you can do. No specialized terms without a definition. No scare tactics.

What This Guide Covers

  • Emergency and “near me” searches have barely changed. Research and comparison content took the real hit.
  • Click-through rates on AI Overview searches fell sharply through 2025, then partially recovered in 2026. Recovery, not restoration.
  • Getting cited inside an AI answer is now the biggest variable you control. Cited pages earn roughly 2.3 times the clicks of uncited pages on the same results page.
  • Your own website is not enough. What other sites say about you carries more weight than what you publish yourself.
  • Ads now appear inside AI answers, and both Google and OpenAI rewrite ad copy by default unless you opt out.
  • Google does not penalize AI-written content, but unedited AI content cannot earn citations, because it contains nothing the AI did not already know.
  • Judge your marketing on leads and booked work, not traffic. Traffic is the metric this shift distorts most.

Part 1: What Actually Changed

For about twenty years, Google worked like a librarian. You asked a question, and it pointed you toward a shelf of books: a list of ten blue links. To get the answer, you had to click.

Two things changed that.

First, Google started answering questions itself. At the top of most searches, you’ll now see an “AI Overview.” Google assembles a written answer from several websites and places it above the regular results. Users get an answer without clicking anything.

How often does this happen? Honestly, nobody agrees, and you should be skeptical of anyone who quotes you one confident number. Published estimates range from under 20% of searches to over 80%, because every study samples different queries. Similarweb reported 43% in 2026, up from 15% a year earlier. Data from Semrush, SE Ranking and seoClarity in late October put it above 80%, including 82.5% of US brand searches.

What is not in dispute is the direction. Google’s own documentation says an AI Overview appears only when its systems judge it more useful than the regular results, so plenty of searches still do not trigger one. But the share has grown every year, and in late August, AI Overviews began expanding automatically into longer responses that look much more like a full AI conversation.

Second, people started skipping Google entirely. Millions of people now type their questions straight into ChatGPT, Google Gemini, Perplexity, or Claude. Those tools give one answer, sometimes with a handful of source links, sometimes with none.

Google is no longer only a librarian pointing at books. It’s increasingly an informed friend who tells you the answer.

That’s good for the person searching. It’s complicated for the businesses that used to get that click.

Part 2: The Three Acronyms, Explained Simply

You’ll hear three terms thrown around. Here’s what each one actually means.

SEO: Search Engine Optimization. Getting your website to show up in the list of links. This is the one you already know. It’s been around for 25 years, and it still matters.

AEO: Answer Engine Optimization. Getting your information chosen as the answer, meaning the text presented in the AI Overview box, the featured snippet, or the response a voice assistant reads aloud.

GEO: Generative Engine Optimization. Getting AI tools to recommend your business by name when someone asks them a question like “who’s the best HVAC company in Kansas City?”

Here’s an analogy that makes the difference clear:

Term

Old-School Equivalent

SEO

Being listed in the phone book

AEO

Being the answer the operator reads aloud when someone calls information

GEO

Being the business your neighbor names when you ask them who to call

Important: These are not three competing strategies where you pick one. They’re three layers built on the same foundation. A business with strong SEO fundamentals is usually the same business that shows up in AI answers, because AI tools read many of the same signals.

Google itself has been unusually direct about this. Its published guidance states there are no additional technical requirements for appearing in AI Overviews or AI Mode. Pages need to be indexed, eligible for Google Search, and genuinely useful to people. The foundations didn’t change. What changed is what gets rewarded on top of them.

Anyone telling you to abandon SEO and buy GEO instead is selling you something.

Part 3: The Most Important Thing in This Whole Article

Here’s the key: how much AI has affected your business depends almost entirely on the kind of questions your customers ask.

The scary statistics you’ve read in the news are real, but they were mostly measured on one kind of question. Let’s break them into three types.

Question Type 1: “I need this right now.”

Examples: “furnace repair near me,” “emergency plumber open now,” “home health agency accepting new patients”

How much has changed: very little.

When someone has an urgent, local, ready-to-buy need, Google still shows the map pack: those three local businesses with reviews, phone numbers, and directions. Research from Whitespark found that on these pure local-intent searches, the map pack appears roughly 93% of the time, while AI Overviews show up only about 15% of the time.

If your business runs on “someone needs us today” searches, the sky is not falling.

Question Type 2: “Help me understand this.”

Examples: “how much does spray foam insulation cost,” “what size furnace do I need,” “signs a parent needs in-home care,” “what does an architect charge”

How much has changed: enormously.

These are the questions AI answers directly. Whitespark’s research found AI Overviews appear on roughly 92% of informational searches. When one appears, fewer people click through to any website.

There’s an important update here, and it runs against the panic narrative. Seer Interactive has tracked this across 53 brands and 5.47 million queries. Click-through rates on AI Overview searches fell hard through 2025, bottoming out at 1.3% in December. Then they reversed. By February 2026, they had climbed back to 2.4%, an 85% recovery in two months. Seer’s own forecasting model had predicted the decline would continue. It didn’t.

The gap today is roughly 2.4% on searches with an AI Overview versus 3.8% on searches without one. That’s a real difference, but it’s closer to a third than the 60% collapse widely reported in 2025.

Read that carefully, though. The click rate recovered, while total click volume kept shrinking. Two things are happening at once. People who only wanted a quick fact now stop at the AI answer, which removes many low-value visits from the math. And Google added more and better-placed links inside its answers. The people still clicking are more serious buyers. There are simply fewer of them.

One honest caveat: this finding is not universally accepted. Plenty of 2026 coverage still reports click-through rates falling by close to 60% when an AI Overview appears. Seer’s recovery data is well documented and drawn from a large sample, but treat it as one credible reading rather than a settled fact.

If you have a blog full of helpful “how it works” articles, this is where your traffic went.

Question Type 3: “Help me choose.”

Examples: “best insulation contractor in Charlotte,” “top-rated home care agency near Tampa,” “architects who design modern farmhouses”

How much has changed: the most. And this is the type that matters most.

These blended searches trigger an AI Overview roughly 97% of the time. This is the moment a customer decides who to call. It used to be a list you could rank in. Now it’s an answer that either mentions you or doesn’t.

This is the single highest-value place to focus.

Let’s make this real with quick examples for a few different industries:

Home services: A local HVAC company in Indianapolis noticed that their “best furnace repair near me” searches are still sending phone calls, but their blog traffic about “what size furnace do I need” dropped sharply as AI Overviews began answering those questions directly. To adapt, they now focus their website content on local success stories and specific advice about Indiana homes, earning more mentions in local trade publications and city guides.

Retail: An independent home goods store saw less traffic from search to their generic “kitchen organization ideas” blog posts. However, their detailed guides like “how to organize small Dallas apartments” started appearing in AI answers and driving branded searches. They now partner with local magazines to provide unique expert tips that mention their store by name.

B2B: A commercial roofing contractor used to rank well for “TPO vs. EPDM roofing,” but those informational searches now get summarized by AI. Instead, they invested in publishing proprietary survey data about roofing failures in their city and got quoted by an industry news website. This mention led to their business being cited by AI tools when users asked “best commercial roofer in Atlanta.”

Notice the pattern: the closer a search is to “I’m ready to buy,” the less AI has disrupted it. The further upstream you go, into research, comparison, and education, the more disrupted it is.

Part 5: What This Means for Your Website Traffic

Your traffic report will look worse than your business actually is. Here’s why. Instead of focusing on overall website traffic, prioritize tracking leads, phone calls, form submissions, and booked jobs. These are the key indicators of how well your marketing is performing, even as traffic numbers shift.

Traffic went down, but the traffic you lost was mostly the traffic that was never going to call you anyway. The person who read your “what is R-value” article and left was rarely a customer. That visit is now happening inside an AI answer instead of on your website.

The visitors who do arrive from AI tools are unusually good. This is the genuinely encouraging finding. Ahrefs published data showing that AI tools sent them only about 0.5% of their visitors, but those visitors accounted for around 12% of their signups. Other 2026 studies have found AI-referred visitors converting several times better than typical search visitors.

The reason is simple. When someone clicks a link inside an AI answer, the AI has already compared the options for them. They’re not starting their research. They’re finishing it.

A fair warning on those numbers: the studies vary wildly, and they’re based on small totals. One early large-scale study found AI traffic converted worse than search, but later studies flipped that finding. Treat AI traffic as a hopeful, fast-growing sliver. For most local and mid-market businesses, it’s currently 1% to 2% of total visits. It is not a channel to bet your budget on this year.

The right way to judge your marketing right now: count leads and jobs, not sessions. If your traffic is down 30% and your booked work is flat or up, your marketing is working fine. If you only look at the traffic chart, you’ll make a bad decision.

Part 6: How a Business Actually Gets Named in an AI Answer

This is the question we get most often, and it deserves its own section: how do we become one of the businesses the AI actually recommends?

First, the number that should shape your whole strategy

If you take one statistic away from this article, take this one.

On the same results page, pages cited inside the AI Overview earn roughly 2.3 times the click-through rate of pages that aren’t cited. Brands cited in an AI Overview see about 35% more organic clicks and 91% more paid clicks.

Getting cited is now the largest single variable you can actually control in search. And it depends far more on how you build your content than on how big your website is, which is exactly why a small local company can beat a national brand at this.

Google just made this more valuable. On October 15, 2026, Google announced it is placing more links inside AI answers and placing them higher. The update adds inline links embedded directly in the words of an AI response, a dedicated section highlighting official websites, expanded news carousels and clickable creator profiles. Read that as a signal: every one of those new link positions goes to a cited source, so Google just raised the payoff on earning citations.

Your website is necessary, but it is not enough.

Start with the good news and the hard news together.

The good news: you cannot buy your way into the recommendation itself. There’s no submission form, no vendor with a special relationship, and no budget that puts your name inside an AI’s answer. Your competitors can’t buy it either. The only way in is to publish information worth citing.

(You can now buy advertising space next to an AI answer. Both Google and ChatGPT sell it. That’s a different thing, and we cover it in Part 7.)

The hard news: your own website, by itself, usually won’t get you there.

This is the finding that surprises people most. Seer Interactive found that brands with active third-party signals appeared in roughly 75% of AI-generated answers, while brands without them appeared in about 1%. Contently’s analysis reached the same conclusion from a different angle: models favor brands that appear consistently across many independent, trusted sources, because repetition across separate publications signals consensus. Self-published content alone rarely carries a brand into AI answers.

Put simply: an AI system trusts what other people say about you more than what you say about yourself.

Mentions count, even without a link.

For twenty years, marketers chased backlinks, meaning clickable links from other websites. That still has some value, but it’s no longer the main event.

AI systems read the web through language, not just links. If a regional trade publication writes “contractors like [Your Company] have reported…” with no link at all, that still teaches the model who you are and what you do. Research has found that branded web mentions, linked or unlinked, correlate closely with visibility in AI Overviews.

This is genuinely good news for smaller businesses, because getting mentioned is much easier than getting linked. Practical places to earn them:

  • Trade and industry association directories
  • Manufacturer and supplier “certified contractor” listings
  • Chamber of commerce and local business organizations
  • Regional trade press and local news
  • Podcast appearances and industry panels
  • Genuine participation in forums and community sites where your customers ask questions

The loop: how expertise turns into phone calls

Here’s the part worth understanding, because it explains why these two goals are really one goal.

Answering “what R-value do I need in an East Tennessee attic” and getting hired to insulate attics in Knoxville are not separate strategies running side by side. The first one causes the second.

Every time your business name appears in credible proximity to your service, your specialty, and your region, it strengthens an association inside these systems. Later, when someone asks “who should I hire to insulate my attic near Knoxville,” the AI reaches for the names most strongly connected to that cluster of ideas.

Your expertise content earns you a place in the consideration set. Your service page is just where they land afterward.

Your unfair advantage: be specific and be local

National websites can publish “what is R-value.” They cannot credibly write about how 1970s crawlspaces in your region behave in August humidity, what your local inspectors actually flag, or what you found in 400 assessments last year.

That content has exactly one credible author: you. And it’s exactly the kind of content that gets cited, because it doesn’t exist anywhere else.

Two practical notes:

  • Numbers and named people outperform adjectives. Research from Princeton found that authoritative quotes and statistics measurably increase how often content gets cited. “IECC Climate Zone 4 requires R-38 to R-49 in attics” beats “proper insulation levels.”
  • Use tables. One 2026 analysis found AI systems extract information from tables at roughly 81% accuracy versus 23% from ordinary paragraphs. A simple cost-range or comparison table is disproportionately likely to be pulled into an answer.

What to stop writing

The generic explainer post is now a losing asset. It gets answered without anyone, and it differentiates you from nobody.

This is a publish less, publish more specifically shift, not a publish-more shift.

Why your report will look wrong

One last warning, because it causes real confusion.

AI visibility usually creates branded demand before it creates trackable referral traffic. Someone reads an answer naming three contractors, remembers the one that sounded credible, and searches that company by name a week later, or types the URL directly.

That lead shows up in your analytics as branded search or direct traffic. The AI answer that actually created it gets no credit.

So the honest scorecard is four things reviewed together, on a quarterly lag: branded search impressions in Search Console, direct traffic trend, AI referral sessions, and your “how did you hear about us?” intake data. Any one of them alone will mislead you.

Here is a simple quarterly tracking routine:

  1. Set aside 30 minutes once every three months to review the four metrics mentioned above.
  2. Open Search Console and note the number of branded search impressions over the past quarter. Record any notable changes.
  3. Check your website analytics for trends in direct traffic and visits referred from AI Assistants.
  4. Gather your intake data. Review how many new leads mentioned finding you through AI, search, or direct referrals.
  5. Jot down a one-sentence takeaway for each metric. Look for patterns or mismatches (for example, if traffic has dropped but leads are steady, your marketing is still working).
  6. Share this recap with your team or advisor.

Doing this quarterly gives you a clear, realistic view of what matters, without getting distracted by week-to-week swings.

Part 7: What This Means for Your Paid Advertising

Two things are happening in paid advertising at once. Google is putting ads into its AI answers, and the AI chatbots themselves have started selling advertising. They’re separate stories, so let’s take them one at a time.

Part A: Google Ads

This is where most articles get it wrong, so it’s worth being precise.

Ads now appear inside and around AI answers. Google confirms that your existing Search, Shopping, and Performance Max ads are already eligible to appear above, below, and within AI Overviews. You didn’t have to do anything for this to start happening.

Ads gained ground on organic. ALM Corp’s analysis found that across every category they measured, organic click share fell 11 to 23 percentage points while text ad click share rose 7 to 13 points. The total pool of clicks shrank, and ads captured a bigger slice of what’s left.

Early reports of collapsing ad clicks have since reversed. 2025 data showed ad click-through rates falling sharply when an AI Overview was present. Seer’s 2026 update found the opposite trend: paid click-through on searches with an AI Overview rose from 14.6% to 16.2% in the first quarter of 2026, while paid click-through on searches without one fell from 26% to 21.8%. Paid clicks are migrating toward the results pages where AI answers appear.

Costs went up. Results also went up. Average cost per click on Google Search rose about 12% year over year into 2026, driven partly by more businesses shifting budget into ads as organic clicks compressed. But WordStream’s 2026 benchmark study of over 13,000 campaigns found something notable: conversion rates improved across 87% of industries, and average cost per lead fell for the first time in five years.

Paying more per click while paying less per lead is a good trade. Judge your ads on cost per lead and cost per job, not cost per click.

The biggest strategic change: your website content now affects your ads. Google has stated that when deciding which ads to show inside an AI Overview, it considers both the search query and the content of the AI Overview itself. Your website content helps shape that Overview.

For twenty years, SEO and paid advertising were separate departments that barely spoke. That’s over. Your content quality now directly influences your ad performance.

One thing to check in your account right now

Google is auto-upgrading older Search and Dynamic Search Ads campaigns to a newer format called AI Max for Search. This is a mandated migration, not an optional feature you can quietly ignore.

If your campaigns were built around tightly controlled keywords and hand-written ads, AI Max trades some of that control for broader automated reach. That can work well. It can also spend your budget on searches you would never have chosen yourself.

AI Max also switches on text customization (previously called automatically created assets) and final URL expansion, which means Google can generate its own headlines and send clicks to pages you did not nominate. You do get controls: the two settings can be turned off separately, URL exclusions keep chosen pages out, generated headlines and URLs appear in your search terms report, and you can remove any asset you do not want.

Ask whoever manages your account three questions: which of our campaigns are scheduled to migrate, is text customization or URL expansion switched on, and how will we measure whether the change helped or hurt? Judge the answer on cost per lead across a full quarter, not on clicks in the first two weeks.

Part B: Advertising Inside ChatGPT

This is new enough that most business owners haven’t seen it yet, and it has progressed faster than almost anyone predicted.

OpenAI began testing ads inside ChatGPT in February 2026 and opened a self-serve platform in May. By the end of August, the company announced that ChatGPT Ads had reached a $1 billion annualized revenue run rate in under 200 days, with tens of thousands of advertisers using it. By late October, availability had reached 89 countries, and OpenAI says ChatGPT now reaches roughly 1.2 billion people each week, though ads still run only on the free and lower-cost tiers. October brought a visual ad format test during image generation in the US, plus a much broader measurement setup: conversion-data integrations, ten attribution partners, geo-lift and brand-lift testing, and brand-safety pilots with DoubleVerify and Integral Ad Science. The original single ad unit has expanded into several formats, and shopping placements now require a product feed.

In short, this went from an experiment to a real advertising channel in about seven months.

The most important thing to understand: these ads appear as a clearly labeled “Sponsored” block underneath ChatGPT’s answer, visually separated from it. OpenAI says ads don’t change the answer itself.

Google uses the same arrangement. You can buy the space beside the answer. You can’t buy your way into the answer. Which is exactly why the work in Part 6 can’t be replaced with an advertising budget.

One caution worth stating plainly: that separation is the platforms’ stated policy, not an independently verified fact. Watch it rather than assume it’s permanent.

Check this setting on both platforms

Both major platforms now rewrite your ad copy by default. At the end of October 2026, OpenAI switched on “creative enhancements” for new and existing self-serve ChatGPT Ads campaigns, which lets its system rewrite headlines, translate ads and change destination URLs without approving each version. Google does something similar through AI Max text customization and final URL expansion. Neither is necessarily bad, and both let you opt out, but they arrive switched on. Ask whoever manages your advertising to confirm in writing which automated rewriting features are enabled on each campaign.

Should your business advertise on ChatGPT?

Maybe. Here’s the honest picture as of September 2026.

Arguments for testing it:

  • There’s no longer a large minimum spend, and the platform now supports conversion-focused bidding rather than just paying for impressions.
  • You can target and exclude by country, region, metro area, and even postal code, which makes it workable for a local service business.
  • The channel is maturing quickly, and being early on a new ad platform has historically been an advantage.

Arguments for waiting:

  • The audience is skewed, and this is the big one. Ads appear only for people on ChatGPT’s free and lowest-cost tiers (”Go”). Paying subscribers on the Plus, Pro, Business, Enterprise, and Education plans see no ads at all. If your customers skew affluent, professional, or corporate, you may be advertising to the group least likely to hire you.
  • The platform changes constantly. Requirements have changed with little notice, including a September change that made product feeds mandatory for shopping placements. Budget time for maintenance, not just setup.
  • Rules for regulated fields are still changing. Category restrictions for healthcare and legal services have been inconsistent. If you’re in one of those fields, check current policy directly before planning around it.
  • It doesn’t work like Google Ads. There’s no keyword bidding. Matching is contextual, based on what the person is actually discussing. Your existing campaign structure and strategy won’t transfer over.

Our current recommendation: this is now worth a genuine test for consumer-facing businesses whose customers are likely using free ChatGPT accounts. For professional services, B2B, and high-end residential work, the audience skew remains a real problem, and we’d still wait. Either way, the organic citation work in Part 6 remains the higher-return investment because it determines whether you show up in the answer people actually trust.

Summary: What to do next by business type regarding ChatGPT Ads:

– Local and consumer-facing businesses: Consider running a small, focused ChatGPT ad campaign to test results if your customers are likely using free accounts. Track leads and interest carefully to see if it’s worth scaling up.

– Professional services, B2B, and high-end residential: Hold off on ChatGPT Ads for now unless you have strong evidence your target audience uses the platform’s free tier. Focus instead on building organic visibility and citations across trusted third-party sites.

– All business types: Prioritize the core steps in Part 6 to increase your chances of organic recommendation inside AI answers. Paid ads in ChatGPT are supplemental, not a replacement.

This lets you quickly identify the next step that fits your business and avoid spending time or budget where the channel isn’t yet mature or well-suited.

Part 8: Your Practical Action Plan

Work through these in order. The early ones are cheap and matter most.

Level 1: Foundations (do these first, they benefit all three acronyms)

  1. Complete every single field in your Google Business Profile. Services, service areas, hours, description, photos, attributes. This is the primary source AI tools use to describe your business. An incomplete profile gives the AI less to work with and less confidence recommending you.
  2. Build reviews steadily, and ask customers to be specific. “Great service” helps you less than “they replaced our 40-year-old attic insulation in North Raleigh in one day.” AI systems read review content, not just the star rating. Recency counts too. A burst of reviews two years ago is worth less than a steady trickle now.
  3. Make sure your business name, address, and phone number are identical everywhere. Your website, Google, Yelp, Apple Maps, Bing, industry directories, and your Facebook page. Marketers call this “NAP consistency,” and its importance has quietly shifted. It used to be about rankings. Now it’s about certainty. An AI system has to

be confident you are one specific, real business before it will recommend you by name. Three different phone numbers across three directories read as ambiguity, and ambiguity gets you left out of the answer.

“But we use call tracking. Doesn’t that break this?”

This is the most common question we get on this point, and it deserves a real answer. Many businesses use call tracking services like CallRail or WhatConverts, which display a different phone number depending on how a visitor arrived at the site. That sounds like the exact opposite of consistency.

The good news: done correctly, call tracking does not hurt you. These services use a technique called dynamic number insertion, where a small piece of code swaps the displayed number in the visitor’s browser after the page loads. Your real number stays in the underlying page code. Search engines and AI crawlers, many of which don’t run that code at all, still see your true number.

Four rules keep it safe:

  1. Pick one canonical number and use it as your single source of truth in your website’s code, your structured data, and every directory listing.
  2. Never put a rotating tracking number in a directory listing. Those numbers are designed to change constantly. They belong on your website, not in your citations.
  3. If you want to track calls from your Google Business Profile, use a single fixed tracking number as the primary and add your real number as a secondary number on the profile. Google supports this, and major call-tracking platforms will set it up this way automatically.
  4. Check your structured data. This is the one people miss. The phone number buried in your website’s schema markup is the version machines trust most. If it was hardcoded to a tracking number during a website build, fix it. Most call tracking tools don’t touch it, but it’s worth verifying rather than assuming.

If your call tracking follows those four rules, keep it. The attribution data is worth far more than the theoretical risk.

  1. Add schema markup to your website. This is invisible code that labels your pages for machines: “this is a service,” “this is a price,” “this is a review.” Your web developer or agency handles this. It’s a one-time job with lasting value.

Level 2: Content That Gets Quoted

  1. Answer the question in the first paragraph. AI systems weigh a page’s opening heavily. If your article about insulation costs spends four paragraphs on your company history before reaching a number, you won’t be quoted. Lead with the answer, then explain.
  2. Publish real specifics, especially numbers. Actual price ranges. Actual timelines. Actual measurements. Vague marketing language (“competitive pricing,” “fast turnaround”) is unquotable. AI tools pull concrete facts.
  3. Quote a named expert, ideally someone at your company. Research from Princeton on generative engine optimization found that adding expert quotes and statistics measurably increased how often content got cited. “Our lead estimator notes that most 1970s-era attics in this region are under-insulated by roughly R-20” is far more citable than an unattributed claim, and naming the person makes it stronger still.
  4. Build a real FAQ page. Not five generic questions. Twenty-five real questions your phone staff actually get, each answered clearly in a few sentences.
  5. Update your existing content instead of only writing new content. AI systems favor recent information. A page refreshed this quarter outperforms an identical page last touched in 2022.

Level 3: Be Mentioned Elsewhere

  1. Get named on other websites. Industry associations, local news, supplier and manufacturer directories, chamber listings, trade publications, local “best of” roundups. AI models build confidence in a business from repeated, consistent mentions across many independent sources. Being talked about elsewhere now matters as much as what’s on your own site.
  2. Describe your business the same way everywhere. If you’re a “spray foam insulation contractor” on your site and an “energy efficiency company” on LinkedIn and a “general contractor” on Yelp, you’ve given AI systems three different businesses to be unsure about.

Level 4: Measure It

  1. Turn on AI traffic reporting. Google Analytics 4 now automatically separates traffic from ChatGPT, Gemini, and Claude into its own “AI Assistant” channel, with no setup required. Note that it only catches visits that arrive with a clean referral tag, roughly 60% to 80% of the real total, so treat the number as a floor.
  2. Check Google Search Console for AI visibility. Search Console now reports AI Overviews and AI Mode as separate segments under Performance, then Search appearance, and this reporting expanded worldwide during 2026. It’s the closest thing available to “how often am I appearing in AI answers.”

Know the limitation before you rely on it. The AI report gives you impressions only. Clicks coming from AI Overviews are counted in your main Performance report, but blended in with ordinary organic clicks, so you can’t separate them out. They aren’t missing. They just aren’t isolatable.

  1. Ask every new lead how they found you, and write it down. This is the least technical and most underrated item on this list. When “I asked ChatGPT, and it mentioned you” starts appearing in your intake notes, you’ll know before any dashboard tells you.

Part 9: Should You Use AI to Write Your Website Content?

Almost every client asks this, usually in a slightly nervous voice. Here’s the straight answer.

Using AI to write content can help or hurt you, and the deciding factor isn’t whether a machine was involved. It’s whether you add anything original afterward.

What Google actually says

Google’s position hasn’t changed in three years: it evaluates content quality, not how it was produced. Google doesn’t penalize AI.

The data backs this up. An Ahrefs study of 600,000 pages found that around 86.5% of top-ranking content shows some AI assistance, with essentially no correlation between AI use and ranking penalties. AI-assisted writing is now normal, not risky.

What Google does punish

What does get enforced, and enforced hard, is something called scaled content abuse: publishing large volumes of low-value pages mainly to capture search traffic. Google’s spam policy applies regardless of how the content was created.

The March 2026 core update named this as a primary target, and the results split sharply:

Approach

Typical outcome

50 to 100 AI-assisted articles with genuine human editing, expertise, and fact-checking

Traffic gains of roughly 30% to 80%

1,000 or more unedited AI articles published for volume

Traffic drops of roughly 40% to 90%

The difference wasn’t AI. It was editorial effort.

Google has since formally extended its spam policies to cover generative AI surfaces. The August 2026 spam update was the first enforcement round under that expanded scope, and a fourth spam update of the year ran from late September into October. Enforcement is running more often than it used to, and the August 2026 spam update was the first enforcement round to run under that expanded scope. The direction is clear enough: the same quality standards now apply whether your page is competing for a blue link or a citation inside an AI answer.

And here’s the risk that should concern a small business most: if a large enough share of your site is judged unhelpful, Google can demote the entire domain. Industry estimates put the danger zone around 30% of your pages, with recovery taking six to twelve months even after you clean it up.

Translation: bolting two hundred thin AI articles onto your website can drag down the service pages that were actually generating your phone calls. The blog doesn’t fail in isolation.

The bigger reason to be careful

Everything above is about risk. But there’s a more important argument, and it connects directly to Part 6.

AI writing is, by its nature, a summary of what’s already common on the internet. To get cited in an AI answer, you need something that isn’t already everywhere.

Ask an AI to write a page about attic insulation in your city, and you’ll get something competent that contains exactly zero information these systems didn’t already have. Nothing in it is worth citing. You’ve paid to publish the average of your competitors.

Compare that to: “We assessed 400 attics in this county last year, and 60% failed on the same three things.”

No AI can generate that sentence. That’s precisely why it gets picked up and repeated.

Unedited AI content quietly defeats the exact goal you’re trying to achieve. You cannot become the distinctive local expert by publishing the consensus.

Where AI genuinely helps

Used well, AI is a real advantage, especially for a busy business owner. The best application is one most people overlook:

Record your expert talking, then let AI turn it into an article.

Twenty minutes with your lead estimator, service manager, or senior designer produces genuine expertise, real specifics, and actual opinions. AI removes the writing labor, which was the actual bottleneck all along. The knowledge is authentically yours. Only the typing was automated.

Other strong uses:

  • Building outlines and structure
  • Producing a rough first draft for a human to react to and correct
  • Reformatting existing information into tables and FAQ sections
  • Writing meta descriptions, image alt text, and headings
  • Tightening and editing human-written drafts
  • Summarizing research so your expert can focus on adding their own view

Two firm limits

  1. Never publish on a topic where nobody at your company has real expertise. Google’s quality guidelines specifically flag misleading authorship and exaggerated credentials. Don’t invent an author, and don’t claim experience you don’t have. If you can’t cover a topic honestly, skip it.
  2. If you’re in healthcare, home care, public health, legal, or financial services, every page needs named human review. In these fields, an AI error isn’t an SEO problem. It’s a liability problem. An incorrect eligibility rule, dosage detail, or legal deadline can cause real harm. Show a real reviewer’s name and credentials on the page. It protects your audience, and it’s also a strong trust signal.

The simplest test

Before publishing anything, ask: “Does this page contain something that could only have come from us?”

If yes, publish it. If no, don’t, regardless of who or what wrote it.

Part 10: How to Spot Bad Advice

This space has attracted many opportunists. Watch for these:

  • “We guarantee placement in AI answers.” Nobody can guarantee this. There is no ad slot to buy and no submission form. Anyone promising it is either misinformed or dishonest.
  • “SEO is dead. Switch your budget to GEO.” Traditional search still drives the large majority of web traffic. AI visibility is an additional layer, not a replacement.
  • “You need to buy backlinks for AI.” Links still have value for traditional rankings, but for AI recommendations, directory consistency, structured data, and genuine third-party mentions carry more weight.
  • Any report that only shows traffic. Traffic is the metric most distorted by this shift. Leads, calls, form fills, and booked work tell the truth.
  • “We’ll publish 50 AI-written articles a month for you.” Volume was the old game. Publishing thin content at scale is now one of the few things Google actively penalizes, and it can pull down your whole site.
  • An agency that cannot tell you which automated features are switched on. Ad platforms increasingly enable automation by default. Your agency should know, without looking it up, whether your ad copy is being rewritten and whether your landing pages are being substituted.
  • Anyone still quoting 2025 statistics as current. This field moves in months, not years. Several widely repeated numbers from last year have already been revised by the researchers who published them.
  • Anyone still quoting 2025 statistics as current. This field moves in months, not years. Several widely repeated numbers from last year have already been revised by the same researchers who published them.

Part 11: The Bottom Line

Search didn’t disappear. The path from question to phone call got shorter, and more of it now happens before anyone reaches your website.

That rewards businesses that are clear, specific, consistent, and well-reviewed, because those are the businesses an AI system can confidently describe and recommend. It punishes businesses that are vague, inconsistent, or invisible outside their own website.

If you were already doing the fundamentals well, you’re in better shape than you think. If you weren’t, the gap will widen faster than it used to.

The window to build an advantage is open right now, while most of your competitors are still forwarding scary headlines to each other.

Have questions about how this affects your specific business? Before reaching out, try this quick self-checklist:

AI Visibility Self-Check:

– When you search for “best [your industry/service] in [your city]” on Google, does your business name appear anywhere in the AI Overview, featured snippets, or local pack?

– Are you being mentioned by name in local news articles, industry directories, or trusted community roundups, with or without a link?

– If you ask ChatGPT or Gemini “Who is a top [your specialty] in [your region]?” does your business get cited?

– In your website analytics, do you see any sessions from AI assistants or increases in branded search volume over the last quarter?

If you answer “no” to any of these, use the action steps in the guide above to close the gap. We’re auditing client accounts for AI visibility as part of our standard reporting. Contact us to see where you currently stand.

Sources referenced

Facebook
Twitter
Email
Print