The Ultimate Guide to Local SEO Keyword Research: How I Actually Do It in the AI Era

TL;DR

This is the complete local SEO keyword research methodology my team and I run for local service business clients across the country. It’s written for business owners and marketers who want to capture the keywords that drive actual phone calls and leads, not just impressions, in both Google and the new AI search engines. Read it end to end for the full method, or skim the highlights below.

• 8 or 9 out of 10 local service businesses get keyword research wrong from day one. The cost of those mistakes is higher than ever in the AI search era.
• Start with your own Google Search Console data. The highest-leverage keyword opportunities are the sleeping giants already in your data, not new keywords you haven’t found yet.
• Build your seed list across three layers: service terms, modifiers, and geography (cities AND neighborhoods).
• Don’t stuff “near me” into your website content. That’s a Google Business Profile optimization play, not a content play.
• Use SE Ranking for volume and difficulty, Google Keyword Planner for free baseline data, and Google Trends for seasonality and geographic validation.
• Most local service businesses only need 20 to 40 keywords to capture 80 percent of their relevant search volume. Pick 3 north star keywords plus 3 to 5 supporting long-tail terms per page.
• Optimize simultaneously for traditional rankings AND AI Overview citations. The signals that earn one increasingly earn the other.

After auditing hundreds of local service business websites over nearly three decades, I can tell you the same thing happens 8 or 9 times out of 10. The keyword research is wrong from day one. Wrong keywords, wrong intent, wrong assumptions about how customers actually search.

The data backs this up. Industry research from Moz on small business SEO shows that the majority of small businesses don’t optimize for local search at all, and most don’t view local SEO as a priority. Recent Ahrefs research found that AI Overviews trigger on roughly 58 percent of question-based queries and 46 percent of longer multi-word queries, with informational searches dominating that exposure almost entirely. The cost of getting your keyword research wrong has never been higher.

Here’s how I think about it. Local keyword research is like picking a destination for your vacation. This is where I want to get to. Once you know the destination, you build everything else around it: the route, the content, the optimization, the supporting work. Unless you know where you want to go, you’re not going to get there. And the destination today is more specific than it has ever been, because the way people search for local services has fundamentally changed.

In this guide, I’m walking you through exactly how my team and I do local SEO keyword research today. Not the textbook version. Not the generic five-step framework every agency copies from every other agency. The actual process my team and I run for local service business clients across Colorado and across the United States, with the same tools you can use yourself, plus a recorded walkthrough where I share my screen and show you the work in real time.

No gatekeeping. Just the real method.

If you already know you won’t have time to run this methodology yourself, that’s a fair conclusion to reach right now. Local keyword research is bread and butter for my team and me. We do it every week for clients across the country. Schedule a call and we’ll take a look at your keywords together. No pressure, no sales pitch. Otherwise, keep reading.

Why Local Keyword Research Is Different Now

I started doing this work in the late 1990s, before Google was the only game in town. Keyword research back then was a fully manual process. There were no tools that gave you exact search volume. No SE Ranking, no Google Keyword Planner, no Google Search Console. You ran searches, took notes, watched what the competition was doing, and used judgment. I’m dating myself a little bit, but I’m genuinely glad I started this early. Living through every era of search means I’ve seen exactly how the work has evolved, and I get to share the tools I grew up with as they exist today.

Here’s what’s different now. People aren’t just typing two or three words into a search box anymore. They’re talking to AI search engines. Claude, Perplexity, ChatGPT, Google Gemini. They’re typing full sentences. Asking follow-up questions. Having actual conversations about what they need. And those AI search engines are answering with specific business recommendations, often before the user clicks anything at all.

This is not a hypothesis. Google’s own search engineering team has confirmed it publicly. On Google’s Search Off the Record podcast, Nikola Todorovic of Google Search Intelligence stated that average query length is growing measurably and that users are uncovering Search’s ability to answer more complex, conversational questions. Independent research from Ahrefs analyzing thousands of ChatGPT conversations found that the average AI conversation runs roughly 350 words. The average Google search runs about 5 words. The same person searches differently depending on which surface they’re on, and the AI surface is where the conversational shift is happening.

That changes the keyword research game in two ways.

First, the keywords themselves are getting longer and more conversational. “Best plumber Boulder” is still valid. But so is “who is the best emergency plumber in Boulder for a burst pipe at 2am.” Both need to be on your radar. The first is what you optimize your page for. The second is what AI search engines need to find when they pull from your page.

Second, where you show up matters more than where you rank. A page ranking #8 in traditional Google might still be cited inside an AI Overview at the top of the page. A page ranking #1 might still be invisible if the AI Overview answers the question without sending the click. Keyword research now has to account for both layers: traditional rankings AND AI citations.

Most of the keyword research advice on the internet hasn’t caught up to this reality. The methodology you’ll find on the first page of Google for “local keyword research” was written for the search landscape of nearly a decade ago. We’re going to do this differently.

The Three Biggest Mistakes I See After Nearly 30 Years

These three keep coming up in audit after audit. If you fix nothing else, fix these.

Mistake 1: Going for the Shiny Objects

When I first got into search engine optimization, I made this exact mistake. People naturally gravitate toward the short-tail, one or two-word keywords with massive search volume. “Plumber.” “Dentist.” “Lawyer.” Those terms have tens of thousands of monthly searches and they look incredibly tempting.

Here’s the problem. Those keywords are targeting nationwide or even international audiences, and the competition is brutal. You’re a local service business in Denver competing against national directory sites with hundreds of thousands of backlinks. You will not win that fight. Not in five years, not in ten.

What I learned, and what I now teach every client and every student at the University of Strasbourg, is that keyword research is different for every business, every situation, every geographic location. Sometimes it’s better to optimize for a search term that has 10, 20, or 50 monthly searches than to chase a term with thousands or tens of thousands. The 50-search term might convert at 30 percent. The 10,000-search term might convert at 0.1 percent. Do the math. Lower volume with higher intent and lower competition wins almost every time for local service businesses.

Mistake 2: Treating Keyword Research as a One-Time Exercise

Local keyword research is not something you do once at the start of an SEO project and check off the list. It’s an ongoing diagnostic process. The keywords your customers actually use shift over time. Voice search has changed phrasing. AI search engines have introduced entirely new patterns. Seasonality affects what gets searched when.

If your keyword research is older than six months, it’s stale. If it’s older than a year, throw it out and start over.

Mistake 3: Optimizing Your Website for “Near Me”

I’m calling this out briefly here and we’ll go deep on it later. Stuffing “near me” into your website page titles, headers, and content does not help you rank for “near me” searches. This is one of the most persistent myths in local SEO and it costs businesses real money in misdirected effort. We’ll fix it in Step 2 of the methodology.

What You’ll Learn in This Walkthrough

This guide pairs with a video walkthrough where I cover the local SEO landscape end to end, including keyword research, Google Business Profile optimization, the AI Overview layer, and what’s working for local service businesses today. Below is an embedded version of a recent local SEO webinar I delivered. We’re also producing a dedicated keyword-research-only walkthrough where I share my screen and demonstrate the entire workflow across all four tools in real time. That recording is in production and will be added to this page as soon as it’s ready.

Here’s exactly what we cover in the dedicated keyword research walkthrough we’re producing:

1. The three biggest mistakes I see local service businesses make with keyword research after nearly 30 years in this industry.
2. How to find sleeping giants in your own Google Search Console data, meaning the keywords you’re already ranking for but not getting clicks from.
3. A live SE Ranking walkthrough showing how to validate search volume and competition for local terms.
4. The free option. Google Keyword Planner, what it gets right and where it falls short.
5. Using Google Trends to confirm seasonal patterns and geographic demand.
6. Why most local service businesses only need 20 to 40 keywords to capture 80 percent of their relevant search volume.
7. The “near me” myth and why optimizing for it on your website is a waste of time. 8. How keyword research actually works in the AI search era.

In the meantime, this local SEO webinar covers the broader strategic context including a substantial section on keyword research:

Step 1: Start With Your Own Google Search Console Data

Before you spend a dollar on a paid keyword tool, you need to mine the keyword data Google is already giving you for free. Google Search Console is where this starts.

Most local service businesses I audit either don’t have GSC set up, or they have it set up and never look at it. That’s the equivalent of leaving cash on the kitchen counter. The data inside GSC tells you exactly which keywords your site is already showing up for, what position you’re ranking in, how many people are seeing those rankings, and how many are actually clicking. All of that is gold for your keyword research.

Here’s what my team and I do when we open GSC for a new local client.

We go to the Performance report. We filter to the last 90 days. We sort the queries by impressions, descending. Then we scan for two things.

The first thing I look for is what we call sleeping giants. These are pages that are already ranking somewhere on page 2 or page 3 (positions 11 to 30) for keywords with meaningful impression volume. The page is doing 80 percent of the work. It just needs the final push to break onto page 1. These are the highest-leverage keyword opportunities in any local business’s portfolio because the page already has authority for the term. Ranking it higher is faster and cheaper than building a new page from scratch.

The second thing I look for is keywords showing high impressions but zero clicks, especially when the page is already ranking in the top 10. That’s a click-through-rate problem, and increasingly it’s an AI Overview problem. Google is answering the user inside the search results page itself, so the click never happens. Different fix, but you find it in the same place.

This very page you’re reading right now started life as a sleeping giant. Tens of thousands of impressions, almost no clicks, average position around 13. We diagnosed it in a Friday SEO Tip and I’m now rebuilding it using the methodology I’m teaching you. The page you’re reading is the methodology in action.

Here’s a real client example of how this plays out.

Boulder Birth and Health is a certified birth center in Boulder, Colorado. They came to us after being acquired and abruptly dropped by a local hospital. Limited budget, urgent need for leads. The first thing we did wasn’t keyword research from scratch. It was opening their Google Search Console and pulling their existing keyword data. Inside that data, we identified low-hanging fruit keywords already showing impressions, including the term “Boulder Colorado acupuncture.” That keyword wasn’t even on their radar when we started. Within months, “Boulder Colorado acupuncture” climbed from position 47 to position 7, and the lead acupuncturist was booked back-to-back. The whole strategy started with reading their own GSC data.

If you don’t have GSC set up yet, do that first. It takes 15 minutes. Verify your domain, wait a week for data to populate, and you’ll have a foundation for every other step in this process.

Step 2: Build Your Seed Keyword List

Once you’ve mined what GSC already gives you, the next step is expanding outward into keywords your site doesn’t yet rank for. This is where you build your seed list.

I think about local seed lists in three layers.

The first layer is your service terms. Whatever your business actually does. Plumbing services. LASIK eye surgery. Personal injury law. Exterior painting. Family dentistry. Each distinct service deserves its own line item, and most businesses I work with have 5 to 15 of these.

The second layer is your modifiers. The descriptive words customers actually type alongside the service term. “Emergency.” “Affordable.” “Best.” “24-hour.” “Licensed.” “Same-day.” These vary by industry. A plumber might emphasize “emergency” and “24-hour.” A dentist might emphasize “family” and “cosmetic.” Pull these from your actual customer conversations, not from a generic list.

The third layer is geography. Your city, your neighborhoods, your service area suburbs. This is where most local businesses stop too soon. They list “Denver” and call it a day. Real local keyword research goes deeper. Cherry Creek. Highlands Ranch. LoDo. Stapleton. Each neighborhood has its own search volume and its own competitive landscape. Phase One Landscapes, a Denver landscape architecture firm we work with, achieved #1 rankings for multiple “Denver landscape design” variations specifically because we created location-specific content targeting Denver neighborhoods. Phone calls increased 45 percent. The neighborhood layer is where the real volume lives.

When you cross all three layers, you get a keyword grid. Service plus modifier plus geography. “Emergency plumber Cherry Creek.” “Affordable family dentist Highlands Ranch.” “Licensed exterior painter Pearl Street Boulder.” Each of those has lower volume than the broad term, but the intent is dramatically higher and the competition is dramatically lower.

One critical reminder. Listen to the language your customers actually use, not the language your industry uses internally. A plumber writes “pipe remediation” on their website. The customer searches “burst pipe repair.” Always go with customer language. If you’re not sure, look at your existing reviews and customer emails. The phrasing customers use to describe their problem is the phrasing they type into Google.

The “Near Me” Myth Most Local Businesses Get Wrong

Time to fix one of the most persistent and expensive mistakes in local SEO.

I see local service businesses cram “near me” into their page titles, headers, and body content thinking it’ll help them rank for searches like “plumber near me” or “dentist near me.” It won’t. Here’s why.

When someone in Boulder types “dentist near me” into Google, the search engine doesn’t go looking for the words “near me” on websites. It uses the searcher’s device location, their phone’s GPS, their search history, and your Google Business Profile to figure out which dentists are physically closest to them. The map pack you see at the top of those results comes entirely from Google Business Profiles, not from website content.

Search Engine Land confirmed this in their recent guide on near-me search optimization. Their words: it is typically unnecessary to add any location-based keywords or similar modifiers to a search like “dentists.” Google already knows you’re looking in a location like San Francisco, and their local and organic algorithms are designed to prioritize results nearest to you when your search is deemed to have a high local intent.

Stop optimizing your website for “near me.” Optimize your Google Business Profile for proximity signals (categories, service area, photos, reviews, GBP posts). Optimize your website for the actual phrases people type, like “family dentist Boulder” or “cosmetic dentist Pearl Street.” That’s where the work pays off.

If you’re stuffing “near me” into your H1, your meta title, or your URL, take it out. Today.

Step 3: SE Ranking for Search Volume and Competition

Once your seed list is built, you need to validate it. Which terms have meaningful search volume? Which are realistically winnable for a business at your authority level? Which should you prioritize first?

This is where I use SE Ranking. I’m a brand ambassador for SE Ranking and an SE Ranking Academy co-instructor, so full disclosure on the relationship. I recommend it because it’s the tool I actually use every day, including with my University of Strasbourg students who get complimentary access through their coursework.

Here’s the workflow I run inside SE Ranking when validating a local keyword list.

I drop each seed keyword into the keyword research tool. I look at three numbers. Monthly search volume. Keyword difficulty. The current SERP composition (who’s ranking, what kind of pages, are there map pack and AI Overview results).

There’s a clean three-letter framework Ahrefs has been teaching that maps almost exactly to how I think about keyword vetting. They call it the BID formula. Business potential, Intent, Difficulty. Business potential asks whether ranking number one for this keyword would actually move the needle for the business. A keyword like “what is acupuncture” has decent volume and low difficulty but no buyer intent. A keyword like “acupuncture clinic Boulder” shows real intent. Always pick keywords that move the needle. Intent asks what the SERP is currently rewarding. If every top result for your target keyword is a YouTube video and you’re trying to rank a service page, it won’t happen. Match the intent or move on. Difficulty is the SE Ranking score plus the SERP composition I’m already pulling. Pass all three checks before you commit to a keyword. Skip any of the three and you’ll waste months chasing the wrong target.

For local service businesses, I’m specifically looking for keywords with 30 to 500 monthly searches and difficulty scores that match my client’s domain authority. A new business with zero authority can’t realistically chase the highest-difficulty terms, but they can absolutely win on neighborhood-level keywords with lower difficulty. As authority grows over six to twelve months, we expand into harder terms.

I also use SE Ranking’s competitive research tool. Drop in a competitor’s domain. Pull their top-ranking keywords. Filter for the local terms. You’ll find keywords your business should be ranking for that you didn’t even know existed. This is where the highest-leverage opportunities live.

For tracked keywords, SE Ranking gives me daily position monitoring with localized SERP data, including a tag for which keywords are triggering AI Overviews. That last piece has become essential. When a keyword shows the AI Overview tag, I know the page targeting that keyword needs to be cite-worthy in addition to being click-worthy.

If you want to follow along, grab a free SE Ranking trial and run your own keyword list through it.

Step 4: Google Keyword Planner (The Free Option)

Not every local business has the budget for a paid keyword research tool right out of the gate. I get it. If that’s where you are, Google Keyword Planner is your free option.

Keyword Planner is part of Google Ads. You don’t have to spend on ads to use it, but you do need to set up an Ads account. Once you’re in, the keyword research feature gives you search volume estimates for any keyword or list of keywords you input.

Here’s the catch most local businesses don’t know about. If you don’t run any ads, Google shows you wide ranges instead of exact volume numbers (something like “100 to 1K monthly searches”). The exact volume data only unlocks when you have an active ad campaign running. At Boulder SEO Marketing, my team and I run very basic Google Ads campaigns specifically so we can access exact search volume inside Keyword Planner. Since we work primarily with local service businesses, having access to precise local search volume data is critical for keyword research. The cost of running a minimal campaign is small. The intelligence advantage from accurate volume data is significant.

Here’s what Keyword Planner gets right. The data comes directly from Google. It’s the same dataset Google Ads uses to estimate your CPC. Volume ranges are reasonably reliable for high-volume keywords.

Here’s where it falls short. Keyword Planner tends to lump similar keywords together into broader buckets, so the volume numbers it shows aren’t keyword-specific. They’re cluster-level estimates. For local service businesses, that’s a problem because the difference between “Denver dentist” (broad) and “family dentist Cherry Creek” (specific) gets blurred. You also don’t get keyword difficulty scores, SERP feature flags, or competitive analysis. It’s a volume tool, not a strategy tool.

My recommendation. Use Keyword Planner to validate that your seed list isn’t completely off base. If a term has zero searches according to Keyword Planner, it probably has very low volume even at a more granular level. But don’t make final priority decisions from Keyword Planner data alone. The cluster-level reporting will mislead you on the long-tail terms that local businesses actually need to win.

Step 5: Google Trends for Validation

The last tool in my workflow is Google Trends, and most local businesses underuse it dramatically.

Google Trends doesn’t give you absolute search volume. It gives you relative interest over time, plus geographic comparison data. For local keyword research, those two features are gold.

The first thing I check in Trends is seasonality. Some local services are highly seasonal. Landscaping. HVAC. Holiday-related services. If “lawn care Denver” peaks in April and dies in November, that’s something you need to know when planning content publication and ad spend. Trends shows you the pattern in 30 seconds.

The second thing is geographic comparison. Drop in two terms. Side by side. See how interest varies by city or region. This is especially useful when you’re trying to decide whether a keyword is worth targeting in your specific market. A term that’s hot in Atlanta might be flat in Denver. Trends tells you that for free.

The third thing I use Trends for is comparing closely related terms. “LASIK Denver” versus “laser eye surgery Denver.” “Plumber” versus “plumbing contractor.” When you can’t tell which phrasing customers prefer, Trends settles it.

Trends is also the cleanest way to spot emerging keywords before they show up in your other tools. New service categories, AI-driven search shifts, regional language evolution. Watch the trend lines for things that are climbing fast.

A quick check-in

If you’re starting to think this is more involved than you have bandwidth for, that’s a perfectly reasonable conclusion. You don’t have to read the rest. Book a call with me and my team and I will run this exact methodology on your business directly, including pulling your Google Search Console data, validating your keyword list inside SE Ranking, and identifying your sleeping giants. Otherwise, the rest of this guide goes deeper into the prioritization and AI search layers that separate local service businesses who win at this from local service businesses who plateau.

Try It Yourself: Local Keyword Opportunity Starter (Coming Soon)

We’re working on building a free interactive tool that gives you a head start on the methodology above without requiring a paid keyword research subscription. Here’s what it’ll do when it launches.

You’ll drop in your service type, your primary city, and 2 or 3 surrounding neighborhoods you serve. The tool will return a structured keyword cluster organized into five tiers. Money keywords (service plus city). High-intent long-tail variations (emergency, affordable, best, 24-hour). Suburb and neighborhood expansion. Question-based informational keywords. AI search visibility keywords designed to surface in AI Overviews and AI search engine recommendations. Each output will include a short “How I’d prioritize these” recommendation in my voice, explaining which tier to attack first based on your inputs.

It’s not a substitute for the full methodology. But it’ll get you 60 percent of the way to a solid seed list in about 60 seconds. Check back in a few weeks. Or drop us a note and we’ll let you know the moment it’ll live.

How I Prioritize When Forced to Choose

Here’s where I have to laugh a little bit.

When prospective clients ask me how many keywords they should optimize for, I know immediately what kind of agency they’ve been talking to before they got to me. The question itself is from a pre-AI SEO playbook. If you’re talking to an agency right now that’s selling you a package for 8 keywords, 10 keywords, 15 keywords, or 20 keywords, make sure to run.

SEO has completely changed since AI entered the picture. Content now needs to target keywords and phrases holistically because search engines and AI search engines have gotten genuinely good at understanding context. A well-written, well-researched page targeting one or two primary keywords will naturally rank for dozens of related long-tail variations because Google’s natural language processing understands what the page is actually about. Pages don’t rank for individual keywords anymore. They rank for topic clusters.

Ahrefs reached the same conclusion in their recent AEO research. When users send prompts to ChatGPT, Perplexity, or Google’s AI Mode, the AI fans each prompt out into many sub-queries the user never typed. The same question phrased ten different ways produces ten different answers with ten different brands mentioned. You cannot optimize for any single one of those sub-queries because they’re synthetic, generated on the fly, and never repeat the same way twice. What you can do is build comprehensive coverage of the topic so that whichever phrasing the AI generates, your content is in the candidate pool. Topic-level coverage replaces keyword-level coverage. That’s the same conclusion I’ve been arguing from a different angle for years.

Here’s how I actually think about prioritization for a local service business.

First, pick three north star keywords for any given piece of content. Not 8. Not 20. Three. These are the keywords the page is fundamentally about. Everything in the page (the title, the H1, the headers, the structure) is built around those three. Pick them carefully.

That said, alongside the three north star keywords, it’s perfectly fine (and often valuable) to identify 3 to 5 supporting long-tail keywords that the page should also speak to naturally. Both Google and the major large language models now evaluate content based on context, semantic relationships, and topical depth, not just exact-match keyword presence. Supporting keywords let you cover the topic comprehensively without diluting the page’s focus. Think of them as backup vocalists. The north star keywords carry the lead. The supporting keywords reinforce the topical signal.

Second, accept that your three north star keywords are educated guesses, not guarantees. Three or four months after the content is live, you’ll have actual ranking data from SE Ranking and Google Search Console. You’ll see what the page is genuinely getting rewarded for. Sometimes it’s exactly what you targeted. More often, it’s adjacent keywords you didn’t fully anticipate. That’s okay. That’s normal.

Third, go back and re-optimize based on what you’re actually ranking for. Minor tweaks. Adjust a section. Add a new heading. Fold in keywords that are showing impressions but not yet ranking on page 1. This is the post-publication re-optimization workflow that separates pages that climb from pages that plateau.

This page you’re reading right now is the live demonstration of that exact workflow. The original version of this page ranked somewhere around position 13 to 15 for a cluster of “local keyword research” terms with thousands of monthly impressions and almost zero clicks. My team and I pulled the Search Console data, identified what the page was already getting impressions for but not converting, watched what the AI Overviews were citing, and rebuilt the entire piece around those data signals plus the methodology I’m teaching you here. If you found this page through a Google or AI search, you found it because the methodology worked. That’s the proof. Same approach we run for clients. Same approach we just ran on ourselves.

That brings me to the bigger insight I’ve earned over nearly three decades of doing this work for local service businesses.

There is a pattern, and once you see it you can’t unsee it. For most local service businesses, there’s a set of roughly 20 to 40 keywords that make up about 80 percent of all the relevant search volume their business should be optimizing for. That’s it. Twenty to forty keywords. Not 200. Not 2,000.

The 20 to 40 break down predictably. A handful of high-intent money keywords (service plus city). A larger set of long-tail variations (service plus modifier plus city). Neighborhood-level expansions (service plus suburb). Question-based informational keywords. A few AI-search-era conversational variants.

If you nail those 20 to 40, you’ve captured the searches that actually matter for your business. Everything beyond that is noise. The agencies pushing 200-keyword packages are either scared, lazy, or selling something that doesn’t exist.

Real Client Examples With Real Numbers

Let me show you what this methodology produces in practice. Five real businesses, five different industries, all using some version of the workflow I just walked you through.

• Maurer Painting is a local Boulder business. Their most crucial keyword is “exterior painter Boulder, Colorado.” We applied Micro SEO Strategies℠ to their target page and got them ranking #1 in the map pack and #1 in the organic blue links for that term. Single keyword, dual #1, the kind of dominance that turns into actual leads on the phone.
• Boulder Birth and Health I covered earlier in Step 1. Acupuncture page rebuilt around “Boulder Colorado acupuncture.” Position 47 to position 7 in months. Lead acupuncturist booked back-to-back. The methodology that drove that result started inside their own Google Search Console.
• Phase One Landscapes in Denver wanted to compete with national landscaping chains for local searches. We created location-specific content targeting Denver neighborhoods, optimized their Google Business Profile with project photos showing Colorado-appropriate landscaping, and built authority through local backlinks. Result: #1 rankings for multiple “Denver landscape design” variations. Organic traffic up 27 percent. Phone calls up 45 percent. They’re now appearing in AI search recommendations.
• 2020 Institute is a vision surgery practice with multiple Colorado locations. They came to us in a competitive medical market where keywords like “LASIK Denver” cost over $50 per click for paid ads. Within four to five months we had them ranking #1 in the map pack and #1 in organic for that term. They’ve reported a 20 percent increase in business and have extended their engagement multiple times. They’re now also appearing in Google AI Overviews for vision correction searches, which is the ideal outcome for AI-era local SEO.
• Sunova Siberians proves this works outside Colorado too. They’re a Seattle-based hypoallergenic Siberian cat breeder who came to us with a professional website but minimal organic traffic. We applied a focused Micro SEO Strategy℠ targeting specific breeding-related keywords, optimized existing pages with strategic precision, and created content demonstrating breeding expertise and Experience, Expertise, Authoritativeness, Trustworthiness signals. Result: 178 percent boost in organic clicks within weeks. Kitten waitlist filled in two weeks.

Five businesses. Five industries. One methodology. The keyword research is the foundation underneath every one of those wins.

The AI Overview Layer for Local Keyword Research

This is the section most local SEO guides don’t have, because they were written before this layer existed.

AI Overviews are Google’s AI-generated answer summaries that now appear at the top of search results for many queries. The frequency varies sharply by query type. Ahrefs analyzed AI Overview presence across the search ecosystem in their recent AEO research and found roughly 21 percent of all keywords trigger an AI Overview, but the number jumps to nearly 58 percent for question-based queries and 46 percent for queries that are seven or more words long. Almost 100 percent of keywords that trigger AI Overviews are informational in intent. When an AI Overview shows up, traditional organic results get pushed down the page. Click-through rates drop. The user often gets their answer without clicking anything.

For local keyword research, this means you’re optimizing for two distinct outcomes simultaneously.

The first is traditional ranking. You still want to rank in the top 10 of organic results for your target keywords. That part hasn’t changed.

The second is AI citation. You want your content to be the source the AI Overview cites when it answers the user’s question. Citation works differently from ranking. AI Overviews tend to favor content that’s structured for direct extraction (clear question-and-answer formats, well-defined headings, scannable paragraphs, specific data points), authored by demonstrable experts (E-E-A-T signals matter heavily here), and published on domains with topical authority. Ahrefs’s research also found that 43.8 percent of all AI-cited pages are listicles, the format that lets AI build consensus across multiple recommended sources. There’s another counterintuitive finding worth knowing: brand mentions across the web correlate more strongly with AI Overview visibility than backlinks do. Ahrefs’s analysis of 75,000 brands put the correlation coefficient at 0.664. That’s a stronger signal than domain rating, referring domains, or any traditional SEO metric. For local service businesses, this is good news. You can compete on AI visibility without a massive backlink profile if you build genuine presence across LinkedIn, YouTube, Reddit, and the local citation ecosystem.

The 2020 Institute example I gave earlier is a working illustration. They rank #1 organically for “LASIK Denver.” They also appear inside Google AI Overviews for vision correction searches. Both layers, same methodology underneath.

Action-First Keywords AI Cannot Replace

Here’s the part of the AI Overview story that most local SEO guides miss completely. AI can summarize informational content. AI cannot use a tool. There’s a category of keywords where the user needs to actually do something interactive, and those keywords are essentially click-protected against AI Overview erosion.

Ahrefs identified eight modifier categories that signal an action-first query. Calculator, checker, generator, tool, template, finder, planner, maker. Search “mortgage calculator,” “ROI calculator,” “schedule template,” “permit checker,” and the AI Overview either disappears or sits on top of a SERP where the click goes to whoever built the actual tool.

For a local service business, this is a real strategic opportunity. “Roof replacement cost calculator Denver.” “HVAC sizing tool for Boulder homes.” “Painting estimate generator.” “Lawn care budget planner.” Each of those is a service-plus-action-modifier-plus-geography keyword that almost no competitor is targeting, and even fewer are actually building tools for. The keyword has lower volume than “Denver roofer” but the click is essentially guaranteed if you rank, because AI cannot answer the query without sending the user to the tool itself.

This is why I’m building the Local Keyword Opportunity Starter you’ll see embedded in this page. It’s not just a content marketing nicety. It’s the structural answer to AI Overview erosion.

The Compounding Effect of AEO Investment

The other thing local businesses get wrong about AI Overviews is treating them as a threat to traffic instead of an opportunity for traffic quality. A recent Duda study analyzed 858,457 websites and found that sites being crawled by AI engines generate 3.2 times more human traffic, 2.7 times more form submissions, and 2.5 times more click-to-call events than sites the AI engines do not crawl. The Duda data also showed that LLM referral traffic to small business sites grew 73 percent year over year.

Translation. AI traffic is small in absolute volume. AI traffic is large in conversion value. Optimizing for AI visibility doesn’t trade off against human traffic. It compounds with it. At BSM, roughly 30 percent of our new clients now tell us they discovered us through ChatGPT, Perplexity, or another LLM. Two years ago that number was zero.

When you’re doing keyword research today, watch for the AI Overview SERP feature flag in SE Ranking. Any keyword showing that flag needs content built for both layers. The good news is that content optimized for AI citation also tends to rank well organically, because the underlying signals are the same: clarity, authority, expertise, structured information, and original insight that doesn’t exist anywhere else on the internet.

The AI-Generated Query Phenomenon

Here’s something I discovered relatively recently and want to share because it’s going to change how you read your Google Search Console data.

A growing percentage of the queries showing up in GSC are not human queries. They’re AI-generated queries. Queries being run by AI agents on behalf of users. Queries being run by third-party AI tools programmatically searching Google. Queries from Perplexity, ChatGPT, and other AI search engines whose backends pull from Google in real time.

How do you spot them? They tend to be unnaturally long. Structurally awkward in ways humans don’t write. Semantically odd. Ten or fifteen words strung together in patterns that look right at first glance but don’t quite match how people actually phrase questions when typing.

For local service business keyword research, the pattern is fairly consistent. The AI-generated queries usually combine a clear location identifier with the service term, then string together additional modifiers in unusual ways. “Best emergency plumber Boulder Colorado for residential pipe repair available evenings and weekends” is the kind of structure I’m talking about.

The implication for your keyword research is significant. When you’re looking at GSC data and you see a long-tail query with hundreds or thousands of impressions but zero clicks, ask yourself if a human would actually type that. If the answer is no, the impressions are likely AI-agent traffic. That’s not bad, exactly. It means your page is being read and considered by AI systems. But it does mean you should focus your click optimization effort on the clearly human queries in your data, not the AI-generated ones.

It also means the underlying truth I mentioned earlier holds even more strongly. Most local service businesses really do have a core set of 20 to 40 keywords that drive 80 percent of their meaningful business. The AI-generated long-tail noise can be informative, but it shouldn’t drive your priority decisions.

Common Mistakes I See

Quick hit list. Five mistakes that come up in nearly every local audit I run, ranked roughly by how much money they cost the business.

The first is what I covered above. Optimizing the website for “near me” instead of optimizing the Google Business Profile for proximity signals.

The second is keyword stuffing the page title. “Best Plumber Denver Colorado | Emergency Plumbing Denver | 24-Hour Plumber Denver CO.” Google flags this as low-quality. Customers tune it out. Pick one strong keyword for the title, write for the human, and let the related variations show up naturally in the body content.

The third is treating location pages as duplicate content factories. If your only difference between your Boulder page and your Denver page is the city name, neither will rank. Each location page needs unique content reflecting that specific market. Different photos. Different testimonials. Different neighborhoods called out. Different local references.

The fourth is ignoring search intent. This is the single most expensive mistake on the list because it produces traffic that looks like success in your analytics but never converts into phone calls or form fills. Every keyword falls into one of four broad intent types, and understanding which is which decides what kind of page you build for it.

• Informational keywords. The searcher is researching, learning, trying to understand. They’re not ready to buy. Examples: “how does LASIK work,” “what causes a leaky faucet,” “is xeriscaping good for Colorado yards.” These keywords belong on educational blog posts and resource pages. They are not service-page keywords.
• Navigational keywords. The searcher already knows what business they want and is using Google as a directory. Examples: “Maurer Painting Boulder,” “2020 Institute LASIK,” “Boulder SEO Marketing contact.” These are usually low volume but extremely high intent. Make sure your business shows up perfectly for your own brand and location combinations.
• Commercial investigation keywords. The searcher is comparing options before buying. Examples: “best LASIK surgeon Denver,” “top-rated dentist Boulder,” “exterior painters Cherry Creek reviews.” These are gold for local service businesses because the searcher is actively shopping. They belong on comparison-format pages, location pages with strong reviews, and service pages with clear differentiators.
• Transactional keywords. The searcher is ready to buy or contact. Examples: “LASIK Denver cost,” “emergency plumber Boulder,” “schedule HVAC inspection Cherry Creek,” “free estimate exterior painting.” These belong on your service pages with prominent calls to action. The phone number, the form, the booking link. Don’t waste these visitors with a long educational article.

There’s also a fifth category specific to local search worth flagging. Local intent keywords that don’t necessarily fit the categories above but signal proximity-based searching. Examples: “open now,” “open weekends,” “same-day service.” These signal urgency and local availability. Optimize your Google Business Profile for these, not your website content.

If you target informational keywords with a service-page format, you’ll get traffic that doesn’t convert. If you target transactional keywords with a long blog post, you’ll lose buyers who wanted a phone number above the fold. Match the page format to the intent. Always.

The fifth and final mistake is not refreshing your keyword research regularly. The keywords your customers used a few years ago are not the keywords they’re using now. Voice search shifted phrasing. AI search engines introduced entirely new patterns. Some of your old target keywords no longer exist as significant search volume. Run the analysis again every year minimum. Twice a year if you’re in a fast-moving industry.

Where to Go From Here

You have two paths from here.

The first is to take this methodology and run it yourself. Open Google Search Console and look for sleeping giants in your existing data. Build your seed list across the three layers. Validate volume in SE Ranking or Google Keyword Planner. Pick your three north star keywords for your most important page. Write something better than what’s currently ranking. Republish, monitor for three to four months, then re-optimize based on real data.

One reality check before you start. Ahrefs’s recent research found that AI citations refresh aggressively. Roughly 45 percent of AI citation sources change every two days. Your keyword research is not a one-time exercise. The terms your customers used six months ago are not the terms they’re using now, and the citation pool AI engines pull from is in constant churn. Whatever you build, plan to revisit it on a quarterly cadence at minimum. Monthly is better for active markets.

If you want the broader local SEO context beyond keyword research, our Local SEO Comprehensive Guide walks through Google Business Profile optimization, AI Overview tactics, and the full Micro SEO Strategies℠ playbook. If you’re earlier in the journey, our 5 Local SEO Best Practices for Small Businesses covers the foundation work. If you’re wondering what professional help costs, our SEO Costs Guide breaks down current pricing models with full transparency.

The second path is to bring us in. If you want a 30-year SEO veteran and his team running this methodology on your business directly, schedule a free strategy call and we’ll take a look at your situation. We’ll pull your GSC data, run a keyword analysis, identify your sleeping giants, and tell you honestly whether SEO is the right channel for your specific business. No pressure. No sales pitch. If we’re not a fit, we’ll tell you that too.

The destination is the same either way. You want to be the local service business your customers actually find when they search for what you do, in Google and in AI search. The keyword research is how you figure out where you’re going. The rest is execution.

Pick your destination. Then go.

Best,
Chris