Introduction
“Furnace repair near me” tells you what someone needs. It doesn’t tell you that their furnace is 18 years old, it’s making a banging noise, another company recommended replacing it, and they want someone nearby who’ll give them an honest second opinion.
Conversational search gives people room to express those details.
Google reports that AI Mode queries are considerably longer than traditional searches, and its own user research found people “intuitively asking longer, more complex questions.” AI Mode uses query fan-out to break multipart questions into related subtopics and searches.
The underlying service is still furnace repair. What changes is how much context the query can carry about the customer’s situation and how they’re likely to choose a provider.
Traditional local keyword research still helps identify demand, but it also needs to account for the recurring circumstances, requirements, concerns, and decision criteria surrounding those searches.
What goes into a local intent map
A local intent map builds on traditional keyword research, which establishes core demand through services, categories, problems, locations, and common modifiers. It adds the customer context that helps determine which business best fits their needs.
For each important service, research these areas:
- Core need: What is the customer trying to accomplish?
- Situation: What caused the search?
- Constraints: What has to be true for a provider to qualify?
- Decision criteria: What will influence which option the customer prefers?
- Geography and timing: Where and when does the need have to be satisfied?
- Decision stage: Is the customer discovering, diagnosing, comparing, validating, or ready to act?
Consider someone searching for a commercial HVAC company. Traditional keyword research might identify “commercial HVAC Boston,” “commercial HVAC contractor,” “commercial HVAC maintenance,” and related terms.
A facilities manager could express the same underlying demand this way:
- “We manage three office buildings around Boston and need an HVAC company that can maintain all three, respond quickly when a system goes down, and handle older rooftop units. Who should we talk to?”
The service is commercial HVAC, but the request also includes several qualifications. The provider needs multi-location capacity, maintenance capabilities, emergency responsiveness, experience with older rooftop units, and commercial specialization.
A local intent map makes those dimensions visible.
| | | | | | | | --- | --- | --- | --- | --- | --- | | Core service | Situation | Constraint | Decision factor | Geography | Stage | | Commercial HVAC | Ongoing maintenance | Three properties | Multi-location capability | Boston | Comparing | | Commercial HVAC | System failure | Rapid response | Emergency availability | Boston | Ready to act | | Commercial HVAC | Aging equipment | Older rooftop units | Relevant system experience | Boston | Evaluating |
This map should sit alongside the keyword list rather than replace it. Keyword data still helps establish the shape of the market. The intent map helps explain what customers within that market are actually trying to solve.
Step 1: Establish the core search demand
Begin with the research you would normally perform. Identify the services people search for, common category terms, geographic variations, important questions, modifiers, competitors, and relative search demand.
- For a dental practice, the initial set might include terms around dental implants, implant dentists, implant cost, sedation dentistry, financing, and geographic variations.
- A managed service provider (MSP) might begin with managed IT services, IT support, cybersecurity services, Microsoft support, and industry-specific terms.
This remains the foundation because conversational search doesn’t eliminate category demand. Someone asking a five-sentence question about an aging furnace still needs furnace repair. Someone describing a complicated IT environment still needs managed IT services.
Keyword data is also useful for prioritization. If one service category has substantially more demand than another, that information should continue influencing the research and content plan. It can also reveal how customers describe services, which may differ from the terminology a business uses internally.
Once those core demand areas are established, the next stage of the research focuses on the context surrounding them.
Step 2: Identify situations and constraints
For every major service, ask what causes someone to look for it and what has to be true for a provider to qualify. The situation describes what’s happening when the search begins.
- For an HVAC company, that might be a system failure, rising energy bills, an unusual noise, aging equipment, or the need to maintain several commercial properties.
- An accountant may see searches triggered by a business sale, multistate tax issues, succession planning, or an IRS notice.
- An MSP may see demand tied to poor support, a security incident, compliance concerns, office expansion, or an upcoming provider change.
Constraints describe the requirements that shape the search.
- A dentist may need to accept a particular insurance plan, provide sedation, accommodate anxious patients, or offer financing.
- An IT provider may need to support Microsoft environments, work with a specific industry, meet compliance requirements, or serve several offices.
Some situations and constraints will appear directly in keyword tools. Others will be easier to find in sales conversations, reviews, intake forms, customer questions, and community discussions.
For each service, ask:
- What would immediately disqualify a provider.
- Which capabilities prospects repeatedly ask about.
- Which requirements come up during sales calls.
- Whether certain technologies, certifications, specialties, accommodations, or geographic needs determine fit.
The goal is to identify recurring patterns rather than every imaginable scenario. If the same situation or requirement keeps appearing across multiple customers, it belongs in the map.
Step 3: Research how customers decide
Once you know what the customer needs and what qualifies a provider, look at how people distinguish between businesses that appear capable of doing the work.
This became especially noticeable during my Ask Maps testing. As queries moved from simple category searches toward situational and trust-oriented questions, Google increasingly framed businesses around qualities such as responsiveness, honesty, specialization, repair-first thinking, and experience with the situation being described.
These decision criteria often explain why one business gets chosen over another, even when both offer the same core service.
- A home services customer might value transparency, workmanship, or responsiveness.
- A healthcare patient might care about bedside manner, appointment availability, or insurance.
- A business choosing a professional service firm might prioritize specialization, industry knowledge, or credentials.
Documenting these factors also helps interpret long-tail demand. A phrase such as “best HVAC company” may indicate quality intent, but the customer’s definition of “best” could involve response time, honesty, commercial expertise, maintenance capability, or experience with specific equipment.
The more clearly those decision factors are understood, the more useful the keyword research becomes.
Step 4: Mine real customer language
Keyword tools are only one source in this process. Some of the most useful intent research comes from places where customers explain their needs in their own words.
Sales calls and intake forms are especially valuable because they capture questions close to the buying decision. Talk to the people answering the phone, qualifying leads, preparing proposals, or conducting consultations. Ask which concerns come up repeatedly and what prospects want to know before moving forward.
Reviews reveal what customers considered important after choosing a provider and experiencing the service.
- If dozens of reviews mention that an HVAC company repaired systems other companies wanted to replace, that pattern says something about how the business is perceived.
- If patients praise a dentist for making them comfortable after years of avoiding treatment, those reviews create an association with anxious patients.
- If clients praise an IT provider for fast response across several offices, they reinforce responsiveness and multi-location capability.
My earlier Ask Maps testing found that review language appeared to play a meaningful role in how businesses were framed for more nuanced local queries. As prompts introduced concerns about honesty, overcharging, replacement decisions, and specialization, Ask Maps increasingly used customer language around those qualities when explaining why a business might fit the request.
Competitor reviews can also reveal customer expectations, frustrations, and gaps. Recurring mentions of response time, communication, surprise pricing, scheduling, or follow-through can uncover decision factors that may not appear clearly in keyword data.
Other sources include Google autocomplete and related searches, People Also Ask, website search data, customer emails, chat logs, Reddit, industry forums, competitor FAQs, and customer support conversations.
Across these sources, repetition matters more than isolated examples. One unusual question may not justify changing a content strategy. The same concern appearing repeatedly across reviews, sales calls, Google results, and community discussions probably deserves attention.
Step 5: Build the intent map around recurring patterns
Once you’ve collected the research, organize it by service rather than leaving it as a long list of observations.
For a dental practice, traditional keyword research may identify dental implants, implant dentist, dental implant cost, dental implants near me, and related searches. Customer research may reveal recurring concerns about missing teeth, failed previous work, bone loss, dental anxiety, financing, treatment time, sedation, and recovery.
Group these patterns around the underlying service to make the relationships easier to see. The same concern may appear across several services. Response time, for example, may matter for managed IT, cybersecurity incidents, and support contracts. Anxiety may affect implants, restorative dentistry, and emergency care.
This also changes how we should think about zero-volume and low-volume queries.
Consider a search such as:
- “I need a dentist near me who is good with anxious adults and offers sedation because I haven’t been to the dentist in years.”
That exact phrase may have little or no measurable search volume, but the underlying needs can still be common. Keyword data may show demand for “dentist near me,” “sedation dentistry,” “fear of dentist,” and “dental anxiety,” while reviews and customer conversations reveal how frequently those needs appear together.
The useful research target is the recurring need beneath the phrasing. Conversational search increases the number of ways those needs can be combined, making it less useful to chase every exact formulation.
Step 6: Validate the map with conversational searches
Once the map is grounded in real customer research, AI Mode, Ask Maps, Gemini, ChatGPT, and other conversational systems can help test it.
Start with a basic service query, then add the situations and constraints you identified.
For example:
- “I need a commercial HVAC company around Boston.”
Then:
- “I need a commercial HVAC company around Boston that can maintain three office buildings.”
Then:
- “We have three office buildings around Boston with older rooftop units. We need one HVAC company that can maintain all of them and respond quickly when a system goes down. Who would be a good fit?”
Pay attention to how the system interprets each version. Look at which businesses appear, how they are described, which capabilities are emphasized, and what evidence or sources seem to support those conclusions.
The purpose of this testing is to see how the dimensions in the intent map affect the way businesses are matched to the request. It can also expose additional factors worth investigating through customer research.
AI-generated responses shouldn’t be treated as demand data. Asking a model to generate 100 possible customer prompts doesn’t prove that customers are actually asking them or that those needs are commercially important. Conversational testing works best after the underlying situations and constraints have been established through stronger sources.
Turn intent research into optimization work
The intent map becomes valuable when it changes what you do next.
Suppose an implant dentist discovers that financing, dental anxiety, sedation, treatment time, candidacy, recovery, and provider experience repeatedly appear in the research. Those findings should influence how the business explains and supports the service.
The primary implant page may need stronger information about candidacy, process, experience, and financing. FAQs may address treatment time and recovery. Supporting content may cover sedation or more complex cases. Case studies and reviews can provide evidence that the practice has worked successfully with patients facing those concerns.
The same process applies to Google Business Profile. Categories and primary services establish what the business does, while service descriptions, attributes, photos, reviews, and other profile information can provide more context about the situations the business is equipped to handle.
Businesses should also make relevant capabilities clear across their websites and profiles, whether that’s an HVAC company’s experience with commercial rooftop units and multi-location accounts or an IT provider’s expertise in healthcare and Microsoft environments.
The research also helps evaluate evidence. For each important constraint or decision factor, ask what supports the business’s claim that it can meet that need.
A local accounting firm may say it provides M&A advisory and tax planning. A case study involving a business sale provides stronger evidence of experience. Team biographies can establish relevant expertise. Reviews, articles, presentations, or third-party coverage can reinforce that association.
If an attribute repeatedly appears in the research but is difficult to verify online, you’ve found a gap worth addressing.
This approach also helps avoid creating hundreds of thin pages for increasingly specific prompt variations. Content should be built around recurring needs and decision problems rather than every sentence someone might type.
Track representative query classes
Conversational search also creates a measurement problem. Once situations, constraints, timing, trust, and customer characteristics can be combined freely, there is no practical keyword list capable of representing every possible query.
A more manageable approach is to track representative classes of intent alongside traditional rankings. Those might include category, service-specific, situational, capability, trust, comparison, and multi-constraint recommendation queries.
An HVAC company could track a category query such as “HVAC company near me,” a rooftop-unit query for service specificity, an aging-system problem for situational intent, a multi-building search for capability, and a second-opinion query for trust. A more detailed recommendation prompt can test several dimensions at once.
Using a consistent sample over time can show which businesses appear, how they are described, what attributes are emphasized, and which evidence sources surface. This adds a conversational layer to traditional rank tracking without pretending an infinite prompt universe can be reduced to a fixed list.
Local keyword research has a larger job to do
Conversational search brings more of the customer’s real-world situation into the query, including problems, requirements, preferences, trust concerns, and decision criteria. Local keyword research needs to account for those recurring circumstances alongside traditional search demand.
For local SEOs, the opportunity is to connect keyword data with customer research. Search data shows where demand exists, while reviews, sales conversations, and customer questions reveal what customers actually need.
Conversational testing can then show how search systems interpret those needs. The future of local keyword research may be less about finding every query people type and more about understanding the recurring reasons behind who they choose.