A solo mental health counselor built a therapist search and rating platform where the AI agent is not a widget in the corner. It is the way the site works: it interviews clients about their therapist, verifies the rating, and helps the next visitor find the right therapist by concern and location.

Finding a therapist is a trust problem, not a search problem

Type "therapist near me" into any search engine and you will get pages of directories. What you will not get is the thing people actually want to know: what is this therapist like to sit across from? Was there empathy? Did clients feel respected? There are plenty of places to search for a therapist. There is almost nowhere to search where the results are tied to what real clients experienced.

Dwayne Stone noticed that gap from an unusual vantage point: he is a therapist himself. A licensed mental health counselor (MS, LMHC) practicing in Seattle, he set out to build Rate My Therapist, a platform where clients rate their therapists on evidence-based questions about the experience, and where anyone can search for a therapist with those ratings attached.

"There's plenty of places to search, but there isn't a place to search where you're actually going to tie it to a rating. That's something uniquely new. To this day, there's not one that I know that's only for therapists."

Dwayne Stone, MS LMHC, Founder, Rate My Therapist

There was one problem. As Dwayne puts it: "I'm a mental health person, I don't do any technology stuff. I struggle with Zoom meetings and how to share a screen."

About Rate My Therapist

Rate My Therapist is a consumer platform founded by Dwayne Stone, who also runs DRS Therapy, his counseling practice in Seattle, Washington. The platform does two things: it collects structured, verified client ratings of therapists, and it lets visitors search for a therapist by name, concern and location, with those ratings attached.

What makes the problem hard is the space it lives in. In behavioral health, users arrive cautious. They are being asked to share how therapy felt, sometimes in a difficult season of their lives, on a website they have never visited. A wall of form fields is exactly the wrong greeting.

The challenge: a sensitive survey, a search engine, and a founder of one

Three requirements pulled against each other:

• The rating had to feel safe, not clinical. The survey asks structured, evidence-based questions about respect and empathy. Presented as a static form, sensitive multi-step surveys bleed users at every field.
• Ratings had to be trustworthy. An open rating site invites noise. Rate My Therapist needed a verification model, so that a rating could be traced to a real client relationship without exposing anyone's identity.
• It all had to be runnable by one non-technical founder. No engineering team, no ops team. Whatever powered the platform had to be built, changed and maintained in ordinary conversations with a partner, not in code.

The solution: one AI agent, both journeys

Working with Dwayne, Tars deployed an AI agent that is, in effect, the platform's interface. Visitors are greeted with one question: would you like to find a therapist, or rate a therapist? Both journeys run as conversations.

Rating as a guided conversation. Instead of a form, the agent interviews the client: overall experience first, then the evidence-based questions, one at a time, in plain language. After the flow was tuned and shortened, 63% of users who start a rating now complete all five questions.

Verified ratings, conversationally. Therapists list themselves through the agent and receive a unique verification code to give their clients. When a client rates, the agent asks for the code, marks the rating verified, and notifies the therapist in real time that a review has come in. No code? The rating still goes through, it is simply not marked verified.

Search by concern, not just by name. A visitor can ask for a therapist by name, or describe what they need: a location and a concern, like anxiety or OCD. The agent searches the platform's curated therapist database first and can fall back to a wider web check for therapists who are not yet listed.

A back office the founder never touches. Every rating and every therapist registration lands in a structured database automatically. What began with third-party connectors now runs natively inside Tars, which means fewer moving parts and nothing for a team of one to babysit.

The deployment did not start this way. It launched in 2024 as a structured, button-driven chatbot. In mid-2025, Tars rebuilt it as a conversational AI agent: natural dialogue, flexible input handling, and smarter search. That upgrade is where the numbers moved.

The difference it made

Engagement nearly doubled. In the eleven months after the AI upgrade, the share of site visitors who engaged with the agent climbed from 27% to 52%. Same site, same audience, same founder: the difference was the conversation.

The survey stopped losing people. Nearly two thirds of users who start a rating finish all five questions. In a category where sensitive multi-page forms routinely lose most of their users, the conversational format is doing quiet, structural work.

The agent revealed what the platform really was. Dwayne built Rate My Therapist to collect ratings. His users had other plans:

"I thought this was going to be a rate-my-therapist thing, and I never anticipated that people were going to use it as a search. I was just amazed at how many more people search as opposed to rate."

Searches outnumber ratings roughly five to one, thousands of them in the last year. At one point, a majority of all site visitors were using the agent specifically to look for a therapist. Because every search is a conversation, the platform knows what people are looking for: the concern, the location, the moment. That insight, surfaced by the agent, reshaped the product's direction.

It feels safe, which in this space is the product.

"Somebody told me it's a very safe-feeling platform, like it's solid. And that's important in my business, in mental health, because people are always worried about confidentiality."

And it never needed a technical team. Over two years of operation, flow changes, question edits, a verification system and a full AI rebuild all happened through working sessions between Dwayne and the Tars team.

"As my website changed and the focus changed, it was always flawless. When we changed the flow, or changed to AI, there was never a hiccup in the integration."

Working with Tars, in the customer's words

Dwayne has rolled out software into mental health settings before, and describes those projects as some of the hardest he has seen. This one, he says, went differently:

"Your quality of work and the collaboration in the team is hard to beat. I've been in a lot of places where we rolled out software for mental health clinics, and it's always a really difficult task. This was a very smooth experience."

Midway through the engagement, he sent the team an unprompted note with the subject line "Hard Working Team," asking how he could send a thank-you gift from Seattle to Bangalore: "I know it's a small project and I still believe in it. I appreciate your team helping me understand and develop it further."

Why this matters for behavioral health

Behavioral health has a front-door problem. The people arriving at a practice's website, a clinic's intake page, or a directory are often anxious, uncertain, and privacy-conscious. Forms greet them like paperwork. Phone lines greet them with hold music. Rate My Therapist shows what happens when the greeting is a conversation instead: engagement doubles, sensitive surveys get finished, and the organization learns what its visitors actually need, question by question.

And it shows something else: none of this required a technical team. A solo, self-described non-technical founder ran a two-journey, verification-backed conversational platform because the AI agent, and the team behind it, carried the technical weight.

As Dwayne sees it, this is where the field is heading:

"I think there's a space for AI-driven mental health help. We're always going to have real therapy, but it's so much easier to talk with AI first and get that additional help. That's where we're going."

For clinics, intake teams, directories and platforms in behavioral health, the pattern is portable: put a safe, structured conversation at the front door, and the front door starts doing intake, triage and discovery on its own.

Want to see what an AI agent could do at your front door? Schedule a demo with Tars.