Deciding to start therapy is the hard part. The next question is smaller and completely practical: do you take my insurance, and what does it cost me?

The agent works from the payer list, plans, and self-pay rates you give it. It asks which plan they carry, then tells them what you can stand behind: in network or out for that plan, what a first session and an initial psychiatric evaluation cost, and what your self-pay option is. Anything that needs a real eligibility check goes to your billing team with the details already collected.

Fewer people disappear at the money question, and the ones who reach a human already know what care will cost.

"Find a therapist" is not one question. It is availability, specialty, modality, and fit at once. Most practice websites answer that with a wall of headshots and a filter nobody uses.

The agent works from your own clinician directory: specialties, modalities, languages, telehealth or in person, current openings, and which prescribers are accepting new patients. It asks the few questions that narrow it, then presents the clinicians who genuinely fit and the times they have. It never promises that a specific clinician is available, and it does not answer a symptom description by naming one.

More people land on a clinician who will still be their clinician in six months.

The decision does not arrive during business hours. It arrives at the end of a bad week, at 1am, on a Sunday. Your front desk is closed and your form promises a callback in two business days.

The agent is the thing that is awake. It has the real conversation on the channel they came in on, and either books the first session or captures a qualified record with the plan, the preference, the urgency, and the source attached. When someone is ready now, your team gets an alert with the full context.

Weekend and after hours demand stops expiring into voicemail, in a category where people simply contact the next practice on the list.

Plenty of people ask two good questions and then stop, to check with a partner or look at the deductible. That is not a lost lead. It is a person mid-decision.

Because the channel is a property of each message and not a separate conversation, the same thread picks up wherever they return: the follow-up email two days later, the WhatsApp message a week after that. The agent already knows what they asked, which plan they carry, and which clinician they were looking at. You can also reach back out by email, SMS, or WhatsApp with the one thing that was missing.

The people closest to starting stop falling out of the funnel.

Every buyer here opens the demo by trying to break it. Name me a therapist for what I have. Tell me if this is depression. Getting a clinician, a plan, or a symptom wrong is a legal problem.

The agent answers from your approved content and nothing else, and declines and hands off when a question falls outside it. No diagnosis, no assessment, no treatment guidance. The risky moments run as deterministic flows you define once: crisis response surfaces the 988 Suicide and Crisis Lifeline, stops the intake questions, and escalates to a live person, at 3am exactly as at 3pm.

One written protocol tells your clinical and legal leads what the agent will say.

For a therapy or psychiatry practice, the conversion event is not a form submission. It is a first session on a specific clinician's calendar.

Booking happens inside the conversation. The agent reads live availability from your clinicians' own Google or Outlook calendars and books against your rules: session length, buffers, minimum notice, working hours, and the person's time zone. Multiple clinicians can be assigned round robin, least busy, or in a priority order you set. Everyone gets a real calendar invite and a confirmation, and the clinician gets the notification with the conversation attached.

The first session is on the calendar while the person is still resolved to go.

Most practices know how many people visited and how many forms came in. Almost nobody can say which campaigns paid for the people who actually started therapy.

One analytics home shows conversations, engagement, escalations, and conversion, every trend compared against the previous period. Lead source attribution breaks results down by campaign, UTM source, and landing page, so you can see which spend produces booked first sessions, not conversation counts. Goals count only when the underlying action actually succeeded. Every number is clickable down to the individual conversation, and the improve view ranks what is going wrong most often.

You walk into a budget review with the first sessions and the campaigns that paid for them.

No therapy practice puts an agent in front of people deciding whether to get help on a vendor's word. Marketing tests it, clinical tries to break it, legal asks what happens on the worst day.

Tars runs that cycle as a product step. Simulated people, driven by personas and goals from your content, hold multi-turn conversations with your actual agent before launch, scored two ways: deterministic code checks and LLM-judge evaluators. Your worst-case inputs, including crisis triggers and clinical boundary rules, become standing tests, and you deploy when the scores clear your threshold.

Your reviewers sign off on what they watched it do, and the agent keeps improving on the questions people actually ask.