To get your managed service provider (MSP) recommended by ChatGPT, make your business facts identical everywhere, publish pages that answer the exact questions buyers ask about your information technology (IT) services in your city, earn mentions on sites the engines already trust, add structured data that matches what the page says, and test a fixed set of buyer prompts every month. OpenAI says ChatGPT ranks search results on multiple factors and that placement is not guaranteed, so nothing here is a switch you flip. These five steps are the parts you control, and the same foundations apply to Gemini and Perplexity.
This guide is the MSP and local version of the question. For how this fits a full program, see NUOPTIMA's generative engine optimization (GEO) service for MSPs.
What Buyers Actually Ask
Buyers rarely ask an artificial intelligence (AI) assistant for "managed IT services." They describe their situation and ask for names:
- "Who are the best managed IT providers in Dallas for a 60 person law firm?"
- "Which MSP near me handles cybersecurity for a manufacturer with two plants?"
- "We are outgrowing our one-person IT team. Who should we talk to in Atlanta?"
- "Is [your company] or [competitor] better for co-managed IT?"
Each one combines a service, a place, a buyer type and often a size. Location matters even when the buyer never types a city. OpenAI's help page says ChatGPT can estimate an approximate location from the user's internet protocol (IP) address and rewrite a prompt into targeted search queries. Its own example turns a question about good restaurants nearby into a search for top restaurants in San Francisco. OpenAI has not published an MSP example, but the pattern suggests a buyer in Houston asking for "IT support near me" gets an answer built from Houston results.
The answer is usually a short list of names, not ten links. If your firm is not one of them, the buyer may never reach your site.
How AI Picks Providers
Here is what OpenAI has published about ChatGPT search:
- ChatGPT sometimes partners with other search providers and sends them rewritten, more specific queries.
- It may share approximate location with third-party providers that supply local information, such as nearby business listings.
- To be eligible, a site must allow OpenAI's OAI-SearchBot crawler, and its host or content delivery network (CDN) must allow OpenAI's published crawler addresses.
- Sites that block OAI-SearchBot will not be shown in ChatGPT search answers. GPTBot is a separate crawler used for model training.
- Results are ranked using multiple factors meant to surface relevant, reliable information, and placement is not guaranteed.
OpenAI does not publish the weights. So the job is to give any engine every reason to trust your firm for a given question: it can reach your pages, it can tell who you are and where you work, other sites agree with what you say, and a page of yours answers the question directly. Google sets a similar floor for AI Overviews and AI Mode in Search: a page has to be indexed and eligible to show a snippet before it can appear as a supporting link. How Gemini, Perplexity and Copilot differ is covered in Perplexity, Gemini and Copilot for MSPs.
Step 1: Clean Up Your Name, Address and Phone
Your entity is how engines recognize your firm as one business across the web. Its core is your name, address and phone (NAP), and for MSPs it is often a mess. Acquisitions leave old brand names live. A legal name ending in limited liability company (LLC) appears on one directory and a trading name on another. The office moved, but the old suite number lives on in a dozen listings. If sources disagree, an engine has less reason to name you with confidence.
| Fact | Common MSP problem | Fix |
|---|---|---|
| Business name | Old brand from an acquisition still listed; legal suffix on some profiles | Pick the name on your signage and website and use it everywhere |
| Address or service area | Previous office, a virtual office, or a different suite on each listing | One current address or a defined service area, written the same way |
| Phone | Tracking numbers, a national helpdesk line, an old main number | One local number on your site and your main profiles |
| What you do | "IT solutions" on one profile, "cybersecurity" on another | The same core services and industries in every description |
Start with the profiles you own: your website footer and contact page, your Google Business Profile (GBP), Bing Places, your LinkedIn company page, and review sites such as Clutch and G2. Google's own guidelines ask for the name you use in the real world, an accurate address or service area, and a phone number under your direct control. Google's guidance for its own AI features also asks site owners to keep Business Profile information up to date. The detail for your GBP is in Google Business Profile for MSPs. If you serve clients on site and have no walk-in office, read service area vs office address before you change anything.
Step 2: Publish Pages for Buyer Questions
Most MSP sites have one "Managed IT Services" page and a generic about page, which gives an engine nothing specific to quote for a law firm in Dallas.
Build the pages your buyers' questions point to:
- One page per core service, with what is included, who it suits and how onboarding works
- City pages only where you genuinely serve clients, with local proof such as named clients or response commitments
- Industry pages for the verticals you already win, written in that buyer's language
- Comparison pages that answer "you or them" honestly; see the MSP comparison page guide
- Case studies with a named client, a baseline, a timeframe and a result
Open each page with a two to four sentence answer to the question it targets, then back it up. Use a narrow claim you can prove. "IT support for small businesses" gives an engine no reason to pick you, because every competitor says it. "Co-managed IT for manufacturers with an in-house IT lead" does, if your case studies support it.
Step 3: Earn Third-Party Mentions
Your own site saying you are the best MSP in Atlanta proves little. Independent sources saying it give the model corroborating evidence. OpenAI has not said how ChatGPT weighs them, but its help page confirms it uses other search providers and third-party local listings, so it is reading the wider web, not just your site.
The mentions that tend to matter for MSPs:
- Reviews spread across Google, Clutch and G2, with clients describing the service and the industry in their own words
- Directory listings that match your NAP exactly; the shortlist is in best directories for MSPs
- Local business press, chamber of commerce listings and association member pages
- Vendor partner pages and joint case studies
- "Best MSPs in [city]" roundups on independent sites, because they mirror how buyers phrase prompts
Earn these. Invented reviews break Google's review policies and put your profile at risk, and paid spots on low-quality list sites add little. This is the slowest step, so start it in the first month rather than after the content is done.
Step 4: Add Schema and llms.txt
Structured data tells an engine, in a fixed format, the facts you cleaned up in step 1. For an MSP that usually means Organization or LocalBusiness markup from schema.org on the home and contact pages, carrying the same name, address, phone and service area, plus frequently asked questions (FAQ) markup where a page has a real questions section. Google's guidance for its AI features says structured data should match the visible text on the page, so markup that contradicts the page does more harm than none.
llms.txt is a proposed plain-text file that points AI tools to your most useful pages. It is cheap to add. Treat it as optional, because none of the OpenAI help pages cited here mention it. The full setup, with examples, is in llms.txt and structured data for MSPs.
Step 5: Measure With a Fixed Prompt Set
OpenAI warns that ChatGPT search results and citations "can be incomplete, outdated, or incorrect," so a single screenshot of your name in an answer proves very little. A fixed set of prompts, run on the same schedule, does.
| Prompt type | Example | What it tells you |
|---|---|---|
| Unbranded city | "Best managed IT provider in Denver for a 50 person company" | Whether you make the shortlist |
| Industry | "IT support for dental practices in Phoenix" | Whether your vertical pages are being used |
| Comparison | "[Your firm] vs [competitor] for co-managed IT" | What engines say about you next to a rival |
| Branded | "What does [your firm] do and who does it serve?" | Whether your entity facts come back correct |
Run the set monthly on ChatGPT, Gemini and Perplexity. Log who gets named, which pages get cited and what is said about you. Track branded and comparison prompts separately from unbranded discovery. A buyer who searches your name may still be comparing you with other firms, not choosing you. Add a separate lead source for AI-assistant referrals in your customer relationship management (CRM) system so pipeline, not just mentions, gets counted. NUOPTIMA's MSP AI visibility rankings show who AI names in each covered city today.
What This Looked Like at Cortavo
Cortavo is an MSP for growing United States businesses of roughly 25 to 250 employees. In May 2025 its average Google position was 42.6 and it had near-zero organic pipeline, in a category where much stronger domains held the generic terms.
The program followed the same order as the steps above. It started with technical fixes and entity and site architecture. It then built city-by-industry guides (Dallas, Houston and Atlanta on one axis; manufacturing, construction, law firms and cloud on the other), plus migration guides for its legacy-IT niche. Authority came next: placements on IT and business publications, with more than 40 earned over the engagement, and reviews spread across Clutch, G2 and Google. Citations were tracked against the same fixed set of buyer prompts every month, across ChatGPT, Gemini, Claude and Perplexity.
By May 2026, 12 months in, Cortavo had 224 AI mentions across its tracked buyer prompts and 4x more AI citations than its closest competitor. Its average Google position had moved to 9.2. A new lead source tagged as an AI-assistant referral appeared in its CRM in December 2025 and has recurred every month since. Over the 12 months, Cortavo's CRM attributed $403,330 in contracted sales to organic and AI search, and 8 new clients closed from organic search.
That is one engagement with its own market, niche and budget, not a forecast for your firm. The month-by-month shape, including the quiet first months, is laid out in how long GEO takes.
Where to Start This Week
Run five unbranded prompts for your main city and service on ChatGPT and note who gets named. Then search your firm's name and list every place your name, address or phone appears differently. Those two lists tell you whether your first problem is consistency, missing pages or missing proof.
If you want the numbers behind it, NUOPTIMA's free MSP lead forecast shows where Google and the AI assistants send IT buyers in your cities today, and estimates the qualified leads you miss while your firm is not among the names they get.
Sources
- https://help.
openai. com/ en/ articles/ 9237897-searching-the-web-with-chatgpt, accessed October 25, 2026. - https://developers.
openai. com/ api/ docs/ bots, accessed October 25, 2026. - https://developers.
google. com/ search/ docs/ appearance/ ai-features, accessed October 25, 2026. - https://support.
google. com/ business/ answer/ 3038177, accessed October 25, 2026. - https://nuoptima.
com/ case/ cortavo-msp-case-study, accessed October 25, 2026.