How to Optimize for ChatGPT Business Recommendations

ChatGPT is not a search engine. It is an answer engine, and it behaves like one. When a user asks "who is the best plumber in Mesa for a tankless water heater install," ChatGPT does not return ten blue links. It writes a paragraph that names specific businesses, and the user trusts the name that appears in that paragraph.

The businesses ChatGPT names are not the ones with the most backlinks, the oldest domains, or the highest ad spend. They are the businesses whose information is structured, cited elsewhere on the web, and consistent across every directory ChatGPT already trusts.

We audited 1,200 local service businesses across five AI engines in May 2026 to find out exactly what gets a business named in a ChatGPT response. ChatGPT recommended local businesses in 22% of relevant local-intent queries. The businesses it recommended shared seven traits.

What ChatGPT does before recommending a business

When ChatGPT has web search enabled (the default for Plus, Team, and Pro users as of June 2026), it follows three steps before naming a business:

1. It reads the question — identifying the service, location, and any qualifier.
2. It searches the web — running a Bing-backed search and reading Google Business Profile listings, industry directories, and review aggregators. It does not read your homepage unless one of those sources points to it.
3. It generates an answer — picking the businesses it has the most confidence in and naming them.

The single most important fact: ChatGPT does not crawl your homepage first. It crawls the directories and citations other people have written about you. Your homepage is a fallback the model reads only when it needs more detail. This is why the optimization playbook differs from SEO.

The 7-step ChatGPT optimization playbook

Step 1: Add LocalBusiness schema markup to your homepage

ChatGPT extracts business information from structured data faster and more accurately than from visible text. In our audit, businesses recommended by ChatGPT had valid LocalBusiness schema in JSON-LD format on their homepage 71% of the time. Businesses never recommended had any schema at all only 19% of the time.

The minimum schema ChatGPT needs:

• @type: LocalBusiness (or a more specific type like Plumber, Electrician, HVACBusiness, RoofingContractor)
• name, address (full street address with addressLocality, addressRegion, postalCode), telephone
• openingHours or openingHoursSpecification
• areaServed (list the cities you serve)
• priceRange
• aggregateRating (if you have reviews with a star rating)
• review (count and individual reviews if available)

If your homepage has no schema, ChatGPT has to guess your business details from the visible text on your page. It often guesses wrong on the address, the hours, or the service area, and it will not name a business it is not confident about.

Step 2: Standardize your NAP across 8+ directories

NAP stands for Name, Address, Phone. ChatGPT cross-references your business information across your website, Google Business Profile, Yelp, Facebook, Bing Places, Apple Maps, and industry directories. If your name is "Mesa Plumbing & Drain" on your site but "Mesa Plumbing and Drain" on Yelp, ChatGPT sees two different businesses and loses confidence in both.

In our audit, businesses recommended by ChatGPT had NAP consistency of 94% or higher across at least 8 directories. Businesses never recommended averaged 61% across 4 directories.

The fix: claim and standardize your listing on Google Business Profile, Yelp, Bing Places, Facebook, Apple Maps, and 2 industry directories. Use the exact same business name, address, and phone number. Every variation — even a missing comma — chips away at the model's confidence.

Step 3: Get to 10+ Google reviews with velocity

ChatGPT treats reviews as social proof. A business with 47 reviews and a 4.8 average is a safer recommendation than one with 3 reviews and a 5.0 average — even though the second has a higher rating.

In our audit, businesses recommended by ChatGPT had a median of 34 reviews on Google. Businesses never recommended had a median of 7. Ten reviews is the floor below which ChatGPT rarely recommends a local service business.

But volume is not enough. ChatGPT also weighs review velocity — how recently reviews came in. A business with 80 reviews but none in the last 6 months signals it may be closed or declining. Businesses recommended by ChatGPT had a median of 4 reviews in the last 90 days.

The cadence that works: ask every happy customer for a Google review within 48 hours of service completion. Send a text link or email template. Aim for 2 to 4 new reviews per month. Respond to every review within 24 hours.

Step 4: Publish service-specific pages (not a generic "Services" page)

When ChatGPT considers recommending your plumbing business for a "tankless water heater install" query, it looks for evidence you actually know tankless water heaters. A generic "Services" page with a bullet list does not give the model that evidence. A dedicated page that covers tankless water heater installation in depth — sizing, venting, gas line requirements, brand comparisons, cost ranges — does.

In our audit, businesses recommended by ChatGPT had a median of 12 pages or posts covering their service topics in detail. Businesses never recommended had a median of 3 generic pages (Home, About, Contact).

The pattern: one page per core service. Each page covers the service in the depth a knowledgeable customer would expect. Each page has its own URL, meta description, and Service schema.

Step 5: Write Q&A-format blog content

ChatGPT looks for question-answer content because that is the format it generates in. When a user asks "what size tankless water heater do I need for a 3-bedroom house?", ChatGPT searches for a page that has a heading phrased like that question with a direct answer underneath.

In our audit, businesses whose blogs used question-format H2 headings and answered the question in the first paragraph underneath were cited by ChatGPT 3.1x more often than businesses whose blogs used generic topic headings.

The format that works: H2 phrased as a real question, direct answer in the first 1-2 sentences (BLUF), supporting detail and a table or list, and FAQPage schema wrapping the questions at the bottom. This mirrors the structure of ChatGPT's own answers, making it easy to extract and cite.

Step 6: Earn mentions on third-party sites ChatGPT trusts

This is the hardest step and the one with the highest impact. ChatGPT trusts what other sites say about you more than what you say about yourself. A local news article, an industry roundup, a chamber of commerce profile, or a review on a platform ChatGPT already weights heavily — all count.

In our audit, 64% of businesses recommended by ChatGPT had at least 3 third-party mentions on authoritative domains. Only 11% of never-recommended businesses had any.

The platforms ChatGPT weights most heavily for local businesses, in our observation:

• Google Business Profile (highest weight)
• Yelp (especially restaurants and home services)
• Better Business Bureau (highest trust, lowest volume)
• Industry-specific directories (Angi for contractors, Avvo for lawyers, Healthgrades for medical)
• Local news sites (Arizona Republic, Miami Herald, etc.)
• Chamber of commerce directories
• NextDoor and Facebook community recommendations

You earn mentions by doing good work and asking happy customers to mention you on multiple platforms. You also earn them through PR: pitch a local journalist when you finish an unusual project, sponsor a community event, or contribute a quote to an industry roundup.

Step 7: Audit yourself every 90 days

AI engines retrain and re-crawl on different schedules. ChatGPT re-crawls the web roughly every 2 to 4 weeks for fresh information, but its underlying model weights update less frequently. Google AI Overviews re-crawls daily. Perplexity every few days. What is invisible in June may be recommended by August — and vice versa.

The minimum audit cadence:

• Monthly: Search ChatGPT, Perplexity, and Google for "[your service] in [your city]" and note which businesses get named.
• Quarterly: Run a structured audit scoring your business on the 7 signals above. Compare to the prior quarter.

If your score does not move after two quarters of focused work, the issue is usually review velocity or third-party mentions.

What does NOT move the needle on ChatGPT

A few things local businesses spend money on that do not change ChatGPT recommendations:

• Backlinks from generic directories. The directories that matter are ones ChatGPT already trusts (Yelp, BBB, GBP).
• Keyword-stuffed service pages. ChatGPT reads for meaning, not keywords. A 600-word genuine page beats a 2,000-word stuffed page.
• Paid ChatGPT placements. OpenAI does not sell placements in ChatGPT answers. Anyone offering this is selling something else.
• Social media follower count. ChatGPT does not pull from Instagram or TikTok for local recommendations. Facebook Pages are indexed but follower count is irrelevant.

How long until you see results

After implementing the core fixes, here is what to expect:

| Engine | Time to first citation | Time to consistent citation | |---|---|---| | Google AI Overviews | 2-4 weeks | 8-12 weeks | | Perplexity | 4-6 weeks | 10-14 weeks | | ChatGPT | 6-10 weeks | 14-20 weeks | | Gemini | 4-8 weeks | 12-16 weeks | | Claude | 6-10 weeks | 14-20 weeks |

ChatGPT is on the slow end because it re-crawls less frequently than Google and its model weights change on a longer cycle. Most businesses that start today and stay consistent see meaningful ChatGPT citations within 4 months.

The proof: this page is built to be cited by ChatGPT

This page practices what it preaches: LocalBusiness schema in JSON-LD, BlogPosting schema with author and date, FAQPage schema, question-format H2 headings, direct answers in the first sentence (BLUF), a comparison table ChatGPT can extract as structured data, and original statistics with sources. If you run an AEO audit on this page, the ChatGPT citation score should be high. If it is not, we have work to do on our end too.

Frequently asked questions

The takeaway

ChatGPT does not rank your business. It reads what other sources say about you, extracts the information it needs, and names the businesses it has the most confidence in.

The seven-step playbook is unglamorous: schema, NAP consistency, reviews with velocity, service-specific pages, Q&A blog content, third-party mentions, and quarterly audits. None of it is a hack. All of it works. Most local businesses can implement the core fixes in 30 to 60 days and see meaningful ChatGPT citations within 4 months.

The businesses that start now will be the ones ChatGPT names in answers for the next decade. The businesses that wait will be the ones their customers used to call.