The outcomes teams like yours see with Tars

Around the clock

Parent inquiries answered when families are actually researching

Intake up front

Packet details and insurance collected inside the conversation

One thread

Web, WhatsApp, SMS, and email in a single conversation

Waitlist warm

Families kept informed from inquiry to start of care

Convert

Reach families while they are still choosing you.

Capture and qualify parent inquiries the moment they reach out

Your late-night inquiry gets answered that night, while the parent is still looking.
The agent asks what your intake coordinator would ask: the child's age, the funding source, workable hours. It also notes whether a diagnostic report exists.
It runs on web, WhatsApp, SMS, and email, all in one thread.
Every inquiry lands in Salesforce or HubSpot as a structured record, transcript and campaign attached.
Ready families are flagged for your team first, so inquiries that used to expire overnight become morning pipeline.

Collect intake and insurance details up front so verification can start immediately

In ABA the admissions funnel is the paperwork, and every step waits on the family.
The agent gathers the plan and member details, the policy holder, and the diagnosing provider.
It captures the availability that decides which of your teams takes the child.
Everything runs as a set sequence, so the same fields come back every time.
Even a parent who stops halfway leaves structured fields in your CRM, so verification starts days earlier.

Keep waitlisted families engaged so they start care with you

Your waitlist is not only yours. Almost every family is sitting on several at once.
Families go with whoever reaches readiness first, and six weeks of silence decides that for you.
Event-triggered campaigns on SMS and WhatsApp carry the news: authorization approved, a documentation gap, availability opening nearer to them.
Because the agent answers in the same thread, a parent can reply and get a real answer.
Anything needing a person reaches your intake team with the full history, and your waitlist stops quietly emptying elsewhere.

Route school, pediatrician, and payer referrals into their own lane

Not everyone who contacts you is a parent. School districts, pediatricians, and payer case managers use the same form.
The agent tells them apart early and collects only what your provider relations team needs to respond.
It routes the case straight to that team and skips parent intake. It never speculates about coverage or authorization outcomes.
Because these conversations touch schools and payers, it hands off to a person quickly.
Your referral relationships stop depending on who checked the inbox, and your parent funnel counts actual parents.

Platform

The platform underneath, built for a conversation your clinical and legal leads have to sign off on.

Control exactly what the agent is allowed to say to a parent

A parent describes their child and asks whether it sounds like autism. That comes early.
The agent stays operational, never clinical. It does not diagnose and does not comment on severity.
It makes no promise about hours, progress, or a start date.
Answers come only from content your team has approved, and outside that the agent says so and hands off instead of improvising from a general model.
Refusals run as set sequences you review as written. Your clinical and legal leads know precisely what a parent will be told.

Book the intake call inside the conversation, while the parent is still there

Your conversion event is a booked intake call, and a separate scheduling page is where that intent dies.
The agent books in the thread, on live availability from your team's own Google or Outlook calendars.
It honors the buffers and working hours you set, and distributes across your coordinators by round robin, least busy, or priority order.
Qualifying details are collected conversationally, then carried into the calendar invite, confirmation email, and coordinator notification.
Ready families land on your intake team's calendar inside the conversation that started the inquiry.

See where families drop off between inquiry and start of care

You know how many inquiries came in and how many children started care. The middle is dark.
Conversion funnels show how many visitors become conversations, how many qualify, and how many complete each step you define.
Field-capture analytics show which details never come back, which is usually where a packet stalls.
Every number opens to the conversation behind it, and any view exports to CSV.
Your marketing lead can name the source that produced starts of care, and your intake lead can see which step is costing families.

Test the agent against your hardest parent questions before a parent ever asks them

Somebody internal has to try to break it first. That review cycle is the real gate.
Before launch, Tars runs simulated parents, driven by personas and goals, against your actual agent.
Every transcript is scored two ways: deterministic code checks, and LLM-judge evaluators on answer quality and boundaries.
Human review validates the evaluators, and you set the thresholds the agent must clear to deploy.
You go live on evidence your own reviewers produced, and the improvement loop keeps running as real conversations teach you what you missed.

How Tars Agents Get Better

The healthcare acquisition flywheel

Putting an agent in front of people asking about care is not a click-a-button decision. Your clinical, legal, and marketing leads all have to sign off first. Tars closes the loop end to end. Train, test, deploy, learn, improve. More inquiries answered and more first appointments booked with every cycle.

Step 1: Train

Connect your service lines, locations and provider directory, accepted payers, pricing and policy pages, intake criteria, and past inquiry history. Set the guardrails and the exact wording for sensitive topics. The agent presents your care the way your best intake coordinator does, from your content and your rules.

Step 2: Test

Run simulated inquiries against the agent before launch, including the ones your reviewers are most worried about. Your escalation protocol and clinical-boundary rules become standing evaluators, so you see how the agent behaves on the hard cases before a single visitor does.

Step 3: Deploy

Go live on web, WhatsApp, SMS, and email when the scores clear your threshold, with booking and CRM write-back switched on from day one.

Step 4: Get Insights

See where inquiries drop off, which campaigns and landing pages produce booked appointments, and how conversion differs by location and service line. Every number opens down to the individual conversation behind it.

Step 5: Improve continuously

Close the gaps, re-test, and raise booked appointments month over month. Each cycle converts more of the traffic you are already paying for.

Create AI Agent

The healthcare acquisition flywheel: train, test, deploy, get insights, improve continuously

Frequently asked questions

What will the agent never say to a parent?

It will never suggest that a child has autism, or that a child does not. It will not comment on symptoms, severity, or behavior, will not give treatment or parenting advice, and will not tell a parent how many hours their child needs or how long progress takes. It will not promise a start date, an authorization outcome, or a place on your schedule. When a parent asks one of these, the agent says plainly that it is a question for your clinical team, and hands the conversation over with everything already gathered so nobody has to start again. Those refusals are set sequences you define and review as written, rather than judgment calls a model makes in the moment, and they are re-tested every time anything on the agent changes.

How do we control what it does say?

The agent answers only from content your team has approved: your service descriptions, your locations and the teams that serve them, your funding sources and accepted plans, your intake requirements, your policies. You control that source list line by line, including removing a page you do not want it learning from. When a question falls outside that content, the agent says so and hands off. It does not fall back to a general model to fill the gap. Before launch you run simulated parent conversations against the real agent, including your own worst-case inputs, scored by code checks and by LLM-judge evaluators with human review validating the scoring, and you set the thresholds that decide whether it deploys at all.

Where does the data land?

In your CRM, as structured fields your team can actually report on rather than a transcript pasted into a notes box. Salesforce and HubSpot are both in production. Each inquiry lands with the child's age band, the location, the funding source, the details the parent provided, whether a diagnostic report exists, the source page and campaign, and the full conversation attached. Partial inquiries land too, so a parent who drops off halfway still leaves you what they had given. Booked intake calls go onto your coordinators' own Google or Outlook calendars. Anything else moves over APIs and webhooks. Tars stores conversations under retention periods you set, with role-based access and a full audit trail.

Is Tars HIPAA compliant, and will you sign a BAA?

Yes to both. Tars is HIPAA compliant with SOC 2 Type 2, ISO 27001, and GDPR, and we sign a Business Associate Agreement before you go live. That matters more here than in most categories, because an ABA inquiry carries a child's age, a diagnosis, and a family's insurance in one conversation, usually on a public page. You get role-based access, retention periods you control, an audit trail per conversation, and explicit control over what the agent can access and store.

Do you have ABA customers we can talk to?

Not yet, and we will not pretend otherwise on a call. Tars has 8+ years in conversational AI and runs patient-facing and family-facing agents in healthcare today, including International Medical Center, Vivant, Avec Group, and Rate My Therapist in behavioral health. What we would rather do than show you a slide is configure an agent on your own intake criteria and your own approved content, and let your team try to break it before you commit to anything.

Does it integrate with the system we run our practice on?

Straight answer: we do not have off-the-shelf connectors to ABA practice-management systems today, and we will not tell you otherwise on a demo. What Tars connects to now is your CRM, Salesforce and HubSpot are both in production, calendar-based booking on Google and Outlook, and APIs and webhooks into the rest of your workflow. Qualified inquiries, intake and insurance details, booked intake calls, and campaign attribution land where your growth and intake teams already work. Direct write-back into a specific ABA platform is a scoping conversation per account. In practice this is the right order anyway, because everything on this page happens before your practice system ever has a record to hold.

We are judged on starts of care, not chat volume. Can you show that?

That is the right way to judge it. Every conversation is attributed: the page and campaign it started from, whether the family qualified, which details were captured and which never came back, whether an intake call was booked, and which location and team they landed with, written into Salesforce or HubSpot as a distinct source instead of collapsing into a generic form fill. The funnel view shows how many visitors became conversations, how many became qualified inquiries, and where the rest stopped, and every number opens to the individual conversation behind it. For a marketing leader defending a budget, that is the difference between a number you can present and an anecdote.

From the first parent inquiry to the start of care, without the drop-off.