Capture the Copay: How AI Front Desks Collect Patient Payments Before the Visit

Twig is an autonomous AI support platform that triages, self-evaluates, and resolves customer support tickets by integrating with tools like Zendesk, Salesforce, and Intercom. For SMB practices — dental, medical, behavioral health, physical therapy, veterinary — Twig handles the text-side reminder, billing, and payment conversations that determine whether a copay or deposit gets captured pre-visit or drifts into accounts receivable. This post is about the highest-ROI accounts-receivable lever an SMB practice can deploy: collecting patient-owed amounts before the patient walks in.

TL;DR: Most SMB practices collect 55–70% of patient-owed amounts at the point of service and the rest drifts into accounts receivable — where 30-day AR turns into 60-day AR turns into write-offs. An AI front desk that captures the copay or deposit at booking (T-7 to T-1) flips that ratio: practices routinely move to 85–95% pre-pay rates, cutting 60+ day AR by 50–70% and writeoffs by 60–80%. This post is the AR mechanics, the per-vertical math, the integration architecture (PMS + payment processor + voice AI + SMS), and the compliance posture for HIPAA-covered practices.

Key takeaways:

• SMB practices typically collect 55–70% of patient-owed at point of service
• AI front desk pre-visit payment ask lifts that to 85–95%
• 60+ day AR drops 50–70%; writeoffs drop 60–80%
• Pre-visit ask at T-7 to T-1 has the best conversion; later asks convert worse
• HIPAA-compliant deployment requires BAA, PCI-DSS payment processor, and minimum-necessary disclosure
• Twig handles SMS and chat-based payment-link sequences; pair with PMS and payment processor integrations

The AR aging curve, in one chart

Practices learn the hard way that delayed collection ages badly. Industry benchmark data on patient-owed AR:

Days outstanding	Probability of full collection
0 (pre-visit or at visit)	95–98%
30 days	78–85%
60 days	55–65%
90 days	35–45%
120+ days	18–28%

The slope from day 0 to day 90 is brutal. Every dollar that doesn't get paid at or before the visit loses roughly 1% of probability per day for the first 90 days. By day 120, you've lost more than 70% of the collectible amount.

The implication: the highest-leverage AR action is moving the collection point from "at checkout" or "after visit" to "before visit." Everything else — collections calls, statements, payment plans — is recovery from a worse starting position.

Why practices don't collect pre-visit (and why that's changing)

Three historical reasons:

1. Front desk capacity — staff doesn't have time to call every patient at T-1 to collect a copay
2. Insurance verification timing — the practice doesn't know the patient-owed amount until 1–3 days before the visit when the payer responds
3. Patient experience concern — "asking for money on the phone feels uncomfortable"

All three are solved in 2026:

• AI front desk handles the T-1 conversation at scale — $0.50/call instead of $4/call human cost
• Real-time eligibility verification (Change Healthcare, Availity, Waystar) returns patient-responsibility amounts within seconds of a booking, hours before T-1
• Patient experience research consistently shows pre-visit pay is preferred over surprise-at-checkout; the discomfort is provider perception, not patient reality

The pre-visit payment sequence

A typical sequence integrated into the broader reminder cadence:

At booking: capture intent + collect deposit (if applicable)

For cosmetic, elective, or high-value bookings, take a refundable deposit immediately ($50–500 depending on procedure). Signals patient commitment, locks the slot, and creates a payment-method-on-file for the rest.

T-7 days: eligibility verification + balance estimate

Behind the scenes, payer eligibility check returns patient responsibility (copay, deductible remaining, coinsurance). Practice gets an estimated amount. Patient gets nothing here.

T-3 days: pre-visit payment ask

Channel: SMS + email. Content:

"Hi Maria — looking forward to your visit Tuesday at 2pm. Your estimated copay is $35. Pay now: [secure link]. This saves you check-in time on Tuesday!"

This is the highest-conversion point — patient is committed, not yet busy with day-of stress, has time to enter card details. Industry conversion: 60–75% pay at this touch.

T-1 day: confirm + payment reminder

Channel: voice AI (the same T-1 confirm call from the no-show sequence). Content:

"Hi Maria, this is the practice confirming your visit tomorrow at 2pm with Dr. Chen. I see your estimated copay is $35 — would you like to take care of that now over the phone, or pay at check-in? I'll send a payment link either way."

Voice AI conversion at this touch: 50–60% of the remaining patients. Combined with the T-3 SMS, this gets you to ~85% pre-pay.

T-2 hours: final friendly nudge

Channel: SMS. Content: "See you at 2pm! Quick reminder: $35 copay due. Pay now or at check-in: [link]". Catches the last 5–8 percentage points.

At check-in: residual

Whatever didn't pay pre-visit pays here. Now down to 10–15% of patients vs. 30–45% pre-AI.

Post-visit: balance billing (rare with proper pre-visit)

The original problem — bills that age into 60-day AR — only applies to the 5–15% who didn't pre-pay AND have remaining balance after the visit. The AR portfolio shrinks dramatically.

The AR math: a representative medical practice

A 3-provider medical practice, 2,000 visits/month, average patient-owed amount $42:

Metric	Pre-AI	With pre-visit AI sequence
Visits/month	2,000	2,000
Total patient-owed	$84,000	$84,000
Collected at/before visit	62%	91%
Collected via post-visit billing	25%	7%
Eventually written off	13%	2%
Days sales outstanding (DSO)	47 days	8 days
Monthly write-offs	$10,920	$1,680
60+ day AR	$14,700	$4,200
Monthly AR recovery from AI deployment	—	$9,240
AI front desk cost / month	—	$600–900
Net monthly AR impact	—	$8,300+

The write-off recapture alone covers the AI front desk cost ~12× over. Plus the DSO improvement is a balance-sheet win — cash that used to take 47 days to convert now takes 8.

Per-vertical adjustments

Dental practices

• Most aggressive pre-pay opportunity: cosmetic and orthodontic ($300–3,000 deposit at booking)
• T-3 SMS for hygiene copays ($20–60); high conversion (75–85%)
• HSA/FSA cards common — payment processor must accept

Medical primary care

• Smaller copays ($15–40) → SMS-only sequence often sufficient
• Eligibility verification critical (Medicare/Medicaid amounts vary)
• Auto-pay-on-file for chronic-care patients reduces touch burden

Behavioral health

• Insurance complexity high; consider self-pay-discount option in the ask
• Lower-key tone — "your session investment is $X" rather than "pay now"
• Higher recurring frequency (weekly/biweekly) → stored card on file is high-value

Physical therapy and chiropractic

• Multi-visit packages — collect package payment at intake, not per-visit
• Cash-pay segment significant; AI handles both insurance + self-pay routing

Veterinary

• Often unique — significant procedures need full pre-pay; routine visits a smaller copay
• Emergency exception: don't gate care behind pre-pay for emergency intake

Architecture: AI + PMS + payment processor + eligibility

Practice Management System (Dentrix / Eaglesoft / Athena / Epic / etc.)
        ↓ (booking event)
Eligibility verification (Change Healthcare / Availity / Waystar)
        ↓ (patient-responsibility amount returned)
AI front desk (Twig for text channels, voice AI vendor for T-1 voice call)
        ↓ (payment ask sequence)
Payment processor (Stripe / Square Healthcare / AuthorizeNet)
        ↓ (charge confirmed)
Webhook to PMS → patient ledger updated, receipt issued
        ↓ (visit happens)
Residual balance (if any) → AI re-engages for follow-up

Three integration questions to ask vendors:

1. Does the AI front desk integrate with your PMS? Common integrations: Dentrix, Eaglesoft, Open Dental, Athena, Epic (via FHIR), eClinicalWorks, ChiroTouch, AvImark. If yours isn't on the list, ask about REST API options.

2. Does the payment processor support your payment mix? HSA/FSA acceptance, ACH for larger amounts, in-house financing for high-end procedures (CareCredit, Sunbit). Some PMSs lock you into their payment processor; check before committing.

3. Does the AI front desk vendor sign a BAA? This is the HIPAA threshold question. No BAA, no deployment in a covered practice.

HIPAA compliance posture

Five must-haves for a covered-entity deployment:

1. Signed Business Associate Agreement between practice and AI front desk vendor. Standard, but verify.
2. PCI-DSS compliant payment processor — card data never touches the AI transcript layer. Use Stripe Payment Links, Square Invoices, or similar tokenized flow.
3. Minimum-necessary disclosure — the AI says "your estimated copay is $35," not "your estimated copay for the wisdom-tooth-extraction procedure with Dr. Chen for diagnosis X is $35." Less detail in the spoken/SMS message is better.
4. PII screening + redaction in transcripts — handled at ingest by the AI vendor's PII screening layer. Twig applies the same pattern on text channels.
5. Audit logging — every payment request, every disclosure, every escalation logged with timestamp and patient ID. Retain per HIPAA + state retention rules.

State-level rules (California CMIA, Texas HB 300, New York SHIELD) add layers on top of HIPAA — verify with counsel if practicing in stricter states.

Why pre-visit conversion is higher than provider expectation

The historical mental model: patients hate being asked for money. The actual data (from healthcare consumer surveys, MGMA, and industry reports):

• 78% of patients prefer knowing the exact amount due before the visit, not at checkout
• 81% are willing to pay online if the link is simple and secure
• 67% find pre-visit payment "less stressful than checkout payment"
• Pre-visit pay correlates with higher net promoter score, not lower

The discomfort with pre-visit payment lives in the provider's office, not in the patient's experience. The deployments that work treat pre-pay as a service to the patient, not a demand on them.

What about the patients who can't pay?

A non-trivial percentage of patients can't pay the full copay pre-visit. The right deployment design handles this:

• Payment plan option in the AI sequence: "I see your copay is $185. Would you like to pay $50 today and the rest after the visit?"
• Financial assistance routing: practices with sliding-scale or assistance programs have the AI offer to connect to the financial counselor instead of demanding payment
• Cash-pay discount: many practices offer 10–20% discount for self-pay; AI surfaces this

These conversations should be common, not exceptional. The AI front desk handles them well when its policy guardrails are configured for empathy + accuracy. The self-evaluation loop Twig uses on text-side prevents the AI from drifting into pressure tactics; the same posture is appropriate on voice.

Where Twig fits

For SMB practices, Twig handles the text-side payment-link sequences — SMS, email, web chat — and pairs with voice AI vendors for the T-1 voice confirm + payment call. Specifically:

• SMS payment-link sequences at T-3, T-2hr, post-visit
• Email balance statements with one-click pay
• Web chat questions about charges, insurance, or "why am I being asked to pay?"
• Post-visit follow-up for any residual balance — autonomous through 60–90 days, then handed to staff for the 5–10% genuinely difficult cases

The shared substrate: PMS as source of truth, payment processor as transaction layer, AI vendors as the conversation engines on their respective channels.

The bottom line

Patient-owed AR is the largest hidden inefficiency in most SMB practice finances. Practices that move collection from "at or after the visit" to "before the visit" — at scale, via AI front desk — flip their AR aging curve, cut writeoffs by 60–80%, and free 40–60% of front-desk-staff time previously spent on balance reminders.

The deployment isn't complicated. It's BAA + PMS integration + eligibility + payment processor + AI sequence — most vendors handle the wiring. The harder part is provider comfort with the change. Once a practice sees the first 30 days of AR aging compress from 47 days to under 10, the comfort issue resolves itself.