
Is ChatGPT Getting Your Therapy Practice Wrong?
Updated: Oct 5
I use ChatGPT all the time. It helps me think through ideas, organize projects, and find clearer ways to explain things in my practice. It had become a normal part of how I worked.
Then one day, I realized I could also use it to help with SEO. I asked it to audit my practice the way a search engine—or a potential client using AI—might see it. I expected a list of keywords, missing links, or pages that needed work. Instead, the audit showed me something much more unsettling: ChatGPT had a surprisingly confused picture of my business. It had old fees, listed services I do not offer, and attached 1,351 unverified facts to my practice. That did not mean every one of those facts was wrong. It meant ChatGPT could not connect them to a reliable source, but it was still using them to form an answer about who I was and what I did.
So I started fixing it. I traced the bad information back to old directories, corrected what I could, cleaned up my own website, and made the important facts about my practice much harder to misunderstand. This guide is the process I ended up using. I wrote it for other clinicians because most of us are paying attention to our websites and Google rankings, but we have not yet thought much about what AI may be telling people before they ever reach our websites.
I am a private-pay couples therapist in Joplin, Missouri. I work exclusively with couples using Emotionally Focused Therapy. I do not accept insurance. I do not offer telehealth. Those are not complicated facts, but AI platforms were getting them wrong, or hedging them, or simply omitting them in ways that made my practice sound like something it is not.
The core problem is this: AI platforms do not call your office. They pull from what exists on the web and treat volume and authority as proxies for truth. If incorrect information appears in enough places, the AI repeats it with confidence.
The good news is that this is fixable. It does not require a technology background. It requires a structured audit, clear documentation, and a deliberate content strategy. Here is what I did, and what I will keep doing.
Step 1: Go Back Through Your Archives and Actually Look
Most practice owners assume their AI presence is neutral or positive because they have never actively checked. I made that assumption too, until I went back through my ChatGPT conversation history and read my own questions and the answers side by side.
If you use ChatGPT regularly, your archive is already there. Go look at it. Search your practice name. Search your city and specialty together. Read what came back.
If you have not been asking ChatGPT about your practice, start now. Use prompts that mirror what a real prospective client would type:
"Who are the best couples therapists in [your city]?"
"Tell me about [your practice name]."
"Does [your practice name] accept insurance?"
"What therapy approaches does [your practice name] use?"
Do the same on Perplexity and in Google's AI Overview. Record every response verbatim. Screenshot it. Date it. You are building an evidence file.
What to Flag
As you review the responses, look for anything in these categories:
Category | Examples to Watch For |
Factual inaccuracies | Wrong fees, wrong session length, wrong location |
Service errors | Insurance listed when you are self-pay; telehealth listed when you are in-person only |
Credential errors | Wrong licensure type, missing credentials, outdated titles |
Sentiment framing | Negative or ambivalent language about your practice or specialty |
Missing information | Core services not mentioned; specialty areas absent entirely |
When I reviewed my own results, the pattern was clear. The AI had absorbed information from outdated directory listings and third-party aggregators that had never been corrected. The 1,351 unverified facts were not all wrong, but they were unanchored, and that ambiguity is where sentiment starts to turn negative.
Step 2: Trace Where the Bad Information Came From
AI platforms do not invent information from nothing. They source it. Once I knew what was wrong, I started working backward to figure out where the AI had picked it up.
For therapy practices, the usual culprits are predictable:
Directory listings that were never updated. Psychology Today, Healthgrades, Zocdoc, and similar platforms often retain outdated information long after a practice changes its fees, drops insurance panels, or moves locations. AI platforms treat these high-authority directories as reliable sources and amplify whatever they say.
Third-party review aggregators. Sites that compile reviews from multiple platforms sometimes introduce errors in how they describe a practice, pulling from old profile versions or misattributing information from a similarly named practice nearby.
Your own old website content. A blog post from three years ago that mentioned a fee you no longer charge, a service you no longer offer, or a credential you have since updated is still indexed and still readable by AI crawlers.
Local business listings. Google Business Profile, Yelp, and Bing Places can carry errors that persist for years if no one is actively managing them.
In my case, several third-party sources were describing Radical Relationship Transformation in ways that conflicted with how I actually operate. That inconsistency is what AI platforms interpret as uncertainty. When the AI cannot confidently confirm a fact, it either omits it or hedges, and hedging reads as negative sentiment to anyone asking a direct question.
Key takeaway: The AI is only as accurate as the sources it trusts. Your job is to become the most authoritative, consistent, and clearly written source about your own practice.
Step 3: Fix the Source Layer Before Adding Anything New
Before I created any new content, I cleaned up what already existed. New content cannot override bad source material if that bad source material is still live and indexed. This step is not glamorous, but it is the foundation.
Work through this checklist:
Log into every directory profile you have and verify that your fee, session length, modality (in-person vs. telehealth), insurance status, and specialty areas are accurate. Update anything that is wrong or missing.
Audit your Google Business Profile. Confirm your address, phone number, hours, and service description are current. Remove any service categories that do not apply to your practice.
Review your own website for outdated content. Check every service page, your about page, and any blog posts that mention fees, credentials, or logistics. Update or remove content that no longer reflects how you actually work.
Search your practice name in quotes on Google and review the top results. If a third-party site is describing your practice incorrectly, contact that site directly to request a correction or removal.
AI platforms re-crawl sources regularly. If you correct the source, the AI will eventually correct its output. If you skip this step and only add new content, you are building on a cracked foundation.
Step 4: Create Content That Leaves No Room for Misinterpretation
Once my existing sources were clean, I focused on giving AI platforms something better to work with. This is where content strategy becomes the most powerful tool available to a private practice owner.
AI platforms favor content that is specific, consistent, and written in plain declarative language. Vague or jargon-heavy descriptions get interpreted loosely. Clear, factual descriptions get cited accurately.
What I Published and Why
A detailed About page that reads like a fact sheet. I rewrote my about page so it states, in plain language: my name and credentials, my licensure, my specialty and what that specialty does and does not include, my session format, my fee, my location, and that I do not accept insurance. I did not bury these facts in narrative prose. I stated them directly, then built the narrative around them.
FAQ content that answers the exact questions AI platforms get asked. I thought about what a prospective client might type into ChatGPT and wrote content that answered each question in one or two clear sentences. This format works because AI platforms are built to extract and repeat direct answers. "Does Radical Relationship Transformation accept insurance? No. This is a private-pay practice." That sentence is hard to misrepresent.
Service pages that are explicit about what I do and do not offer. I work exclusively with couples. I use Emotionally Focused Therapy, not CBT or DBT. Sessions are 90 minutes. I named these distinctions explicitly on every relevant page. Specificity is what separates content that gets cited accurately from content that gets interpreted loosely.
The Consistency Rule
Whatever you publish on your website needs to match what is in your directories, your Google Business Profile, and your Psychology Today listing. Inconsistency across sources is one of the primary signals that causes AI platforms to generate hedged or negative sentiment. When every source says the same thing in similar terms, the AI has no reason to doubt it.
Step 5: Keep Checking, Because This Is Not a One-Time Fix
Correcting your AI presence is not a project you finish and close. AI platforms update their outputs as they encounter new information, and new information appears constantly in the form of new reviews, updated directory listings, and newly indexed web pages.
I set a quarterly reminder to run the same audit queries I ran in Step 1. I compare the current responses to my baseline screenshots and look for:
New errors that were not present before
Corrections that have taken effect (this is worth documenting; it confirms the strategy is working)
New services or credentials that are not yet reflected in AI responses
This is where most practices fall short. They do the initial audit, make corrections, and assume the problem is solved. But AI outputs shift over time. A practice that was accurately described in January may have drifted by July if new third-party content has introduced conflicting information.
I now track my AI presence alongside my traditional SEO metrics. I treat AI accuracy as a standard part of practice reputation management, not a one-time technical fix. The clients who find me through AI platforms made a decision based on what those platforms said about me before they ever visited my website. What those platforms say needs to be accurate, and keeping it accurate is ongoing work.
Going back through my ChatGPT archives was uncomfortable. Seeing 1,351 unverified facts attached to my practice name, and reading responses that did not reflect how I actually work, was not fun. But it gave me something concrete to fix. If you have not looked, look. The information is already out there. The only question is whether it is accurate.
You may also find useful
If you are ready to get started, Schedule a Couples Therapy Appointment.
Call or text me at 417-768-9089, or email me at shira@radicalrelationshiptransformation.com.

