# Use ChatGPT or Claude to Build a Local SEO Plan That Out-Ranks Bigger Competitors

> Give ChatGPT or Claude your services, local conditions, and buyer questions. Turn the output into a knowledge graph, publish answer-first pages, and become the source AI search wants to cite.

## Knowledge map

- **Type:** Article
- **Canonical page:** https://amtechai.com/articles/build-local-seo-plan-with-chatgpt
- **Format:** Agent-optimized Markdown twin of the HTML article (clean, no page chrome).
- **Topics:** Local authority, Knowledge graph SEO, Local SEO, AI Overviews, Query fan-out, Information gain, ChatGPT, Claude, Local service businesses, Topical authority, strategy
- **See also:**
  - [AI learned to trade stocks before it could flip a burger: automate the operations brain, not the front desk](https://amtechai.com/okf/playbooks/ai-learned-to-trade-stocks-before-it-could-flip-a-burger-automate-the-operations-brain-not-the-front-desk.md) — General operational authority
  - [The $2M bike shop hiding in a town like Salisbury: what AI takes off the owner's plate first](https://amtechai.com/okf/playbooks/the-2m-bike-shop-hiding-in-a-town-like-salisbury-what-ai-takes-off-the-owner-s-plate-first.md) — General topic by market
  - [Phoenix runs on pools and panels: the operational AI a Valley service business actually needs](https://amtechai.com/okf/playbooks/phoenix-runs-on-pools-and-panels-the-operational-ai-a-valley-service-business-actually-needs.md) — General topic by market
  - [Building Sherman: the back-office AI a supply yard needs as the subdivisions land](https://amtechai.com/okf/playbooks/building-sherman-the-back-office-ai-a-supply-yard-needs-as-the-subdivisions-land.md) — General topic by market
  - [AI Overviews](https://amtechai.com/okf/entities/concept-ai-overviews.md) — Tool

## Article

## Answer

Use ChatGPT or Claude to map your business into a local knowledge graph: services, customer problems, materials, local conditions, locations, proof, and buying questions. Then publish one answer-first page for each high-value intersection. Bigger competitors usually have authority. You win by having the most specific, useful, locally true answer on the subquestions AI search now breaks apart.

> **✅ This is not theory for us.**
>
> AMTECH used this exact strategy to push a brand-new, roughly two-month-old site with almost no article library into Google AI Overviews. Not with 10,000 posts. Not with a decade-old domain. With a tighter graph, clearer answers, and pages that made the machine’s job easier.

## Search is turning into a research assistant. That is the crack in the wall.

The old SEO game was simple to explain and miserable to play: buy links, publish keyword pages, wait years, and hope the domain-authority giants do not notice you. That world is not gone. But AI Overviews and AI search changed the shape of the opportunity.

Modern search does not only look for one perfect page for one keyword. Google documents that AI features can use query fan-out: the system breaks a question into related searches, explores subtopics, and assembles a useful answer. That means a local business can win by owning the subquestions the big sites answer lazily.

So stop asking, “Can I rank for landscaping Phoenix?” Ask, “When the machine breaks Phoenix landscaping into heat, caliche, monsoon runoff, drip irrigation, turf, HOA rules, plant survival, and buyer intent, where do I have the best answer on the internet?” That is a very different game.

## The pyramid structure for every page

| Layer | What goes there | Why it works |
| --- | --- | --- |
| Opening hook | A sharp claim, field result, or buyer pain in the first screen. | Mobile readers decide fast. AI systems also need a clear page purpose. |
| Immediate payoff | A direct answer in the first 80 words, before the story, caveats, or sales pitch. | The page becomes easy to quote, summarize, and trust. |
| Expand | Diagnostic table, examples, instructions, warnings, local constraints, and proof. | This creates topical depth without burying the answer. |
| Expand again | FAQs, internal links, citations, schema, and related next steps. | The page becomes a node in a larger graph instead of an isolated blog post. |

## Turn your business into an entity graph.

An entity graph is just a map of the nouns your business actually touches and the relationships between them: services, problems, materials, places, tools, seasons, regulations, customer types, and outcomes. Search engines think in entities and relationships. AI assistants do too. Your SEO plan should match that structure.

Do not start with keywords. Keywords are symptoms. Start with the world your customer lives in. A Phoenix landscaper does not only sell “landscaping.” They solve dying lawns in July, caliche planting failures, HOA notices, drip leaks, gravel washout, and water-bill panic in specific neighborhoods under specific conditions.

# Examples

### Prompt 1: build the local entity graph

Paste this into ChatGPT or Claude, then replace the bracketed fields.

```text
You are a local SEO strategist. I run a [business type] in [city / region].

Build my local entity graph as an indented tree.

Include:
- services
- customer problems
- materials, surfaces, tools, and assets involved
- local conditions that change the answer: climate, soil, regulations, neighborhoods, housing stock, seasons, water rules, permit rules, HOA rules
- cities, neighborhoods, and service-area modifiers

Then connect them. Show which problems relate to which services, materials, local conditions, and places.

Finally, mark the 15 intersections where a local expert could answer better than a national website.
```

## Example: the graph starts turning into pages

| Intersection | Article question | Why a national site struggles |
| --- | --- | --- |
| Problem × local condition | Why is my new lawn dying in July in Phoenix? | Generic advice says water more. Local advice explains timing, evaporation, heat stress, grass type, exposure, and when turf is the wrong call. |
| Material × place | Can I plant straight into caliche soil in the East Valley? | Generic planting guides ignore the cement-hard layer that traps water and changes root behavior. |
| Problem × season | Why does my gravel wash into the street during monsoon? | Generic articles say add edging. A local answer explains sheet flow, yard pitch, granite size, swales, and low-point control. |
| Service × regulation | What landscaping changes trigger HOA issues in Scottsdale? | National content cannot safely guess neighborhood-level standards, approval patterns, and documentation expectations. |

## Find the gaps authority sites left open.

The graph gives you topics. The gap audit tells you which topics deserve pages. This is where the strategy starts making money, because not every article is worth publishing.

Look for information gain: what can you add that the current results do not? A better diagnosis. A local rule. A photo-backed example. A decision threshold. A mistake warning. A price factor. A reason the obvious answer fails in your city.

# Examples

### Prompt 2: rank the content gaps

This prompt forces the model to separate generic filler from local expertise.

```text
Using the entity graph above, find the pages my competitors probably have not written well.

For each opportunity, give me:
1. The exact question a buyer would type into Google, ChatGPT, Claude, Perplexity, or Gemini.
2. The generic answer a high-authority national site would give.
3. The local expert answer that would add information gain.
4. The proof I need before publishing: photos, job notes, local source, regulation page, customer quote, measurement, or field observation.
5. The internal links this page should include.

Rank the opportunities by information gain, buyer intent, and how hard the answer is for a national competitor to fake.
```

> **⚠️ Important: AI is the strategist, not the witness.**
>
> ChatGPT can structure the plan. It cannot truthfully claim your customer’s yard has caliche, your city requires a permit, or your team saw a specific failure mode unless you verify it. The proof layer is what separates authority from AI slop.

## Turn the graph into a publishing plan.

Now build the schedule. Each page should answer one real customer question, connect to one service, connect to at least one place or local condition, and link to the next logical step. That is how a site becomes wiki-like without becoming bloated.

The point is not “publish a ton of content.” The point is to publish a ton of useful edges in the graph. Problem to service. Service to place. Place to regulation. Material to failure mode. Question to proof. Page to page.

# Examples

### Prompt 3: build the 90-day plan

Use this after the graph and gap audit. It creates an action-first queue instead of a vague content calendar.

```text
Turn the highest-opportunity SEO gaps into a 90-day publishing plan.

Use this page format for every article:
- title as a real customer question
- answer in the first 80 words
- diagnostic table
- step-by-step fix or decision path
- what not to do
- local reason this matters
- DIY vs. call-a-pro threshold
- internal links to service, city, and related guide pages
- visible FAQ
- schema recommendations that match only what is visible on the page

Prioritize pages that create a wiki-like cluster around my services, locations, customer problems, and proof.
```

## The article template that actually deserves to rank

- Open with the direct answer. No “in today’s digital landscape.” No throat clearing.
- Add a diagnostic table so the reader can find their exact situation on a phone.
- Break the instructions by condition, surface, customer type, season, or risk level.
- Include a “what not to do” section because people often search right before making the expensive mistake.
- Explain the local reason the answer changes: soil, heat, water rules, storm behavior, housing stock, permitting, or local buyer expectations.
- Give an honest DIY vs. call-a-pro threshold. Trust converts better than a desperate hard sell.
- Link internally to the service page, the location page, and the related guide that naturally expands the answer.
- Use schema only for what is visible: Article, FAQPage when the FAQ exists, BreadcrumbList, and accurate entity markup.

## The research backs the instinct: useful graphs beat isolated posts.

Knowledge graphs are useful because they represent entities and relationships, not just strings of text. Academic surveys describe knowledge graphs as a way to structure information for search, question answering, recommendation, and reasoning. That is basically the job your website now has: help machines and people understand what you know, where it applies, and why it should be trusted.

Google’s public guidance points in the same direction from a practical SEO angle. AI features can explore related subtopics. Structured data helps clarify what the page is about when it reflects visible content. Helpful content should be original, useful, reliable, and made for people. Local ranking still depends on relevance, distance, and prominence. None of that rewards a thousand thin city pages. It rewards a site that is easy to understand and hard to fake.

Our hypothesis is simple: a local website built like a useful wiki around real services, real places, real questions, and real proof is easier for AI search to retrieve, summarize, and cite than a generic authority site with shallow content. The graph is not a hack. It is the architecture of being useful.

## Aggressive only works if the standard is strict.

Here is the part people skip. The strategy is not hard anymore. You can copy the prompts on this page. The hard part is executing 40 times a quarter without letting quality collapse.

Every page needs research. Every local claim needs verification. Every answer needs to be direct. Every table needs to help. Every internal link needs a reason. And some drafts need to die because they do not add anything. Ten genuinely useful pages beat 100 that all say “call a professional.”

That editorial layer used to require a researcher, writer, strategist, editor, and SEO operator. That is why the good version of this was trapped inside expensive retainers. AI did not remove the need for standards. It removed the excuse for moving slowly.

## This is the system AMTECH is building for operators.

The articles in this library are not random blog posts. They are nodes in a graph: AI employees, local authority, estimating, prompt workflows, owner bottlenecks, service businesses, implementation, pricing, and proof. Each page has a job. Each page links to the next decision.

That is what we build for clients too: the graph, the page standard, the research loop, the publishing queue, the internal links, the schema, and the agent workflow that keeps it moving. The goal is not to rent attention forever. The goal is to become the answer in your category and your market.

## FAQ

### Will ChatGPT or Claude make up local facts?

Sometimes. Use the model for structure, questions, outlines, and first drafts. Verify local claims with official sources, field photos, job notes, customer conversations, and real operator experience before publishing.

### How many pages do I need?

Fewer, sharper pages beat a big pile of thin pages. Start with 20 high-information-gain pages that each answer one real local question better than the current results.

### Do I still need a Google Business Profile and reviews?

Yes. The website explains what you know. Your Business Profile, reviews, photos, citations, and local links help prove you exist, serve the area, and deserve trust.

### Is this just publishing a lot of AI content?

No. Volume without original detail is spam with better grammar. This strategy only works when every page adds a useful distinction: a local condition, a decision threshold, a diagnostic table, a verified source, or a field observation competitors missed.

## Citations

[1] [AI features and query fan-out in Search](https://developers.google.com/search/docs/appearance/ai-features) — Google Search Central
[2] [Helpful, reliable, people-first content](https://developers.google.com/search/docs/fundamentals/creating-helpful-content) — Google Search Central
[3] [Structured data introduction](https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data) — Google Search Central
[4] [Local ranking: relevance, distance, and prominence](https://support.google.com/business/answer/7091) — Google Business Profile Help
[5] [Contextual estimation of link information gain](https://patents.google.com/patent/US20200349181A1/en) — Google Patents
[6] [A survey on knowledge graphs: representation, acquisition, and applications](https://arxiv.org/abs/2002.00388) — arXiv

---

Alternate views: [HTML](https://amtechai.com/articles/build-local-seo-plan-with-chatgpt) · [OKF concept](https://amtechai.com/okf/articles/build-local-seo-plan-with-chatgpt.md) · [llms.txt](https://amtechai.com/llms.txt)