Most clinics measure a new website tool in quarters. This one had numbers worth writing about in 25 days.

An autism diagnostic and ABA therapy clinic — anonymized here, serving families across its region — put an UpChat AI website agent on the front of its site as the first point of contact. No front-desk staff monitoring the widget. No intake coordinator on standby after hours. Just the AI, trained on the clinic's services, age ranges, and intake process, handling every conversation end to end.

We exported the full log from the first 25 days: 63 conversations between June 29 and July 23, 2026. Here's what actually happened — including the parents who showed up at 5 A.M.

The results at a glance

In its first 25 days live, an AI website agent handled 63 conversations for an autism and ABA therapy clinic and captured contact details from 40 of them — an 84% capture rate on conversations the agent flagged as prospective patients. More than half of all activity (52%) happened outside standard clinic hours, when no one was at the front desk.

  • 63 conversations handled in 25 days, with no staff monitoring the chat
  • 40 leads captured (email or phone) — a 63% overall capture rate
  • 84% capture rate on new-patient conversations
  • 52% of conversations arrived outside clinic hours — nights, early mornings, and weekends
  • Nearly half of captured leads (18 of 40) came in after hours
  • 7 recruiting inquiries triaged away from intake, keeping the queue clean

That's roughly 40 qualified contacts a month, without a coordinator touching the chat.

Who's asking — and what they want

Parents looking for autism support don't arrive casually. They show up mid-worry, often after a pediatrician's comment or a long night of research, and they want to know two things fast: can you help my child, and how do we start.

Across 63 conversations, the demand sorted into clear lanes:

  • Diagnostic evaluations led the way. Requests for autism testing — for toddlers, school-age children, and teens — were the single most common reason families reached out.
  • ABA therapy intake was close behind. Roughly a third of conversations involved enrolling a child with an existing diagnosis, with goals from communication and life skills to school readiness and behavior support.
  • Real logistics, too. Families asked about insurance and Medicaid acceptance, wait times, in-home versus in-clinic therapy, and school support — the questions that decide whether a family books or bounces.
  • The full age range showed up. From infants flagged for early intervention to school-age children, teens, and even a set of twins needing evaluation.

The agent didn't just answer these questions. It worked each one toward a next step: asking the child's age, whether there was already a diagnosis or an evaluation was needed, confirming the family was in the service area, and collecting contact details so the intake team could follow up.

The 5 A.M. parent

Here's the part no front-desk schedule solves.

At 5:01 on a Wednesday morning, a parent opened the chat describing a young child's sensory sensitivities and speech delays and asked about getting an evaluation. The agent asked the right early questions, confirmed the clinic could help, and captured the parent's email and phone — hours before the office opened, without waking anyone up.

That parent wasn't unusual. The pattern runs through the whole dataset:

  • On a Sunday at 5:23 A.M., another parent asked about insurance acceptance and ABA plus school support for a child with a prior diagnosis — and left full contact details.
  • On a Saturday at 11:23 P.M., a parent of twins asked about diagnostic testing for behavioral and communication concerns, and left a phone number and email.
  • Just before midnight, a parent of a nonverbal young child asked for communication strategies; the agent talked through AAC and sign-language options and captured their contact for follow-up.

The math holds across the log: 52% of all conversations (33 of 63) happened outside standard clinic hours, and nearly half of captured leads (18 of 40) arrived after hours. A parent who works up the nerve to ask for help at 11 P.M. and hits a "leave a message" box is a parent who keeps searching — and often books with the next clinic that answers.

Keeping the intake queue clean

Capturing leads is only half the job. Just as valuable for a clinic: the agent kept non-patient traffic out of the intake pipeline.

  • Job seekers. Seven conversations were people asking about RBT and behavior-technician roles — hiring interest, not patients. The agent pointed them toward the careers path instead of cluttering the new-patient queue.
  • Out-of-scope requests. About half a dozen adults sought evaluations for themselves or an adult relative. Rather than starting a pediatric intake that would only end in a dead end, the agent referred them to appropriate adult evaluation resources.
  • Wrong service entirely. One visitor was looking for an unrelated service; the agent recognized the mismatch and redirected them.

None of those is a lost lead — they were never the clinic's patient to begin with. But every one is a phone call or email an overworked front desk didn't have to field. A well-trained agent qualifies and deflects, so the humans spend their time on families who fit.

What made the capture rate work

  1. The agent asks for contact info at the right moment. Not on message one — after the parent has described their child, confirmed the age and diagnosis status, and heard that the clinic can help.
  2. Qualifying questions do double duty. The child's age, existing diagnosis or evaluation, insurance, and location all move the conversation forward and hand the intake team a pre-qualified record.
  3. There's always a next step. New-patient requests get routed to intake with context, out-of-scope families get pointed to the right resource, and job seekers get sent to hiring. Nobody is left at a dead end.

What this means if you run a clinic or practice

Autism and ABA clinics, pediatric therapy practices, dental and specialty offices — the pattern is the same. Your prospective patients research at night, reach out in moments of worry, and expect an answer the moment they ask.

  1. Your after-hours demand is bigger than you think. More than half of this clinic's conversations — and nearly half its captured leads — arrived outside office hours.
  2. Specificity builds trust. The agent knew an evaluation request from an ABA intake, asked about age and diagnosis status, and handled insurance questions. Families share contact details with a chat that clearly knows the practice.
  3. You don't need a bigger front desk. This entire dataset was generated with no staff monitoring the widget.
  4. Triage protects your team. Filtering out recruiting traffic and out-of-scope requests kept the intake queue focused on families the clinic could actually serve — the same qualification discipline that helps a Washington leak detection company capture 99% of its serious inquiries.

Frequently asked questions

How many new patients can an AI chat agent capture for a healthcare clinic?
In this deployment, an autism and ABA therapy clinic captured 40 leads from 63 website conversations in its first 25 days — roughly 40 qualified contacts a month. On conversations the agent flagged as prospective patients, it captured contact details 84% of the time.

Does an AI agent work outside office hours?
Yes — and that's where much of the value is. Here, 52% of conversations happened outside standard clinic hours, and nearly half of captured leads arrived on nights, early mornings, or weekends.

Can an AI agent handle sensitive healthcare inquiries appropriately?
The agent's job is to listen, answer service questions, qualify the inquiry, and route it to the human intake team — not to diagnose or give clinical advice. Out-of-scope requests were referred to appropriate resources rather than pushed into intake.

How long does it take to set up an AI website agent?
Most UpChat clinics go live in under five minutes: sign up, train the agent on your website content, customize the widget, and paste one embed snippet.

Your website is already getting the traffic. The only question is whether anyone's there when a worried parent types "do you test young children for autism?" at 5 A.M. See how it works or start your free trial — 14 days, no credit card, live in under 5 minutes.

Methodology: figures come from a full export of 63 website conversations handled by an UpChat AI agent for a single autism diagnostic and ABA therapy clinic between June 29 and July 23, 2026 — its first 25 days live. The clinic anonymized and redacted the data before analysis; no personal or health information is included. Lead capture means the visitor provided an email or phone number. Clinic hours are defined as Monday–Friday, 8 A.M.–6 P.M. local; conversations outside that window or on weekends count as after hours. Examples are lightly edited and fully anonymized. Because the dataset spans under a month, these figures describe early-deployment capture and triage — not enrolled-patient outcomes.

BM

Written by

Brad Holly, MBA
Founder & CEO

Brad Holly, MBA writes about AI chat agents, customer conversations, and what actually moves the needle for growing businesses.

Put your best employee to work

Deploy an AI agent that replies in zero seconds, captures leads, and books demos — 24/7. Live in under five minutes.

Never miss a playbook

One useful email a month on AI, support & growth. No fluff.