At hour four, the patient has already WhatsApp-messaged two other clinics and received a quote from one of them. This is not a hypothetical. When I audited the intake system for a mid-tier Istanbul dental clinic in late 2025: I’ll call them Clinic D: I found that 63% of their inbound leads were receiving their first coordinator response more than four hours after initial contact. Their Meta advertising spend was €4,200/month. Their consultation booking rate was 14%. I’ve built intake systems for clinics across hair transplant, dental, and cosmetic surgery, and Clinic D’s TFCR (Time to First Coordinator Response) was one of the worst I’d seen for a clinic at their volume. The revenue implication was immediate and calculable.
Last Updated: 20260609T0
9 min read
A detailed case study of an anonymized Istanbul dental clinic that deployed the EKSENAI intake system to reduce Time to First Coordinator Response from 4+ hours to under 4 minutes, with specific week-by-week implementation timeline and before/after metrics.
Clinic D had three coordinators, a Chatwoot instance they’d barely configured, and a workflow in n8n that someone had started building and abandoned. They had 80 leads per month and were booking 11–12 consultations from them. Eight weeks after implementation, they were booking 24–27 consultations from the same 80 leads, on the same ad spend.
What Did Clinic D’s Intake System Look Like Before the Intervention?
The baseline audit produced a clear picture of a broken intake pipeline.
| Metric | Before (Baseline) | After (Week 8) | Change |
|---|---|---|---|
| Avg. TFCR (all channels) | 4h 17min | 3.8 min | -98.5% |
| Lead-to-consultation booking rate | 14% | 31% | +121% |
| Consultations booked/month | 11–12 | 24–27 | +118% |
| Coordinator messages per lead (pre-consult) | 22 avg. | 9 avg. | -59% |
| % leads receiving response outside business hours | 0% | 100% (automated) | — |
| Deposit conversion rate (consult to deposit) | 21% | 34% | +62% |
| Estimated monthly Revenue Leakage (recovered) | — | €18,400–26,800 | — |
The Lead Latency problem was structural. The three coordinators managed all incoming WhatsApp messages from a shared number, with no routing logic, no automation, and no visibility into which leads had been responded to. The Coordinator Black Box was literal: there was no way to tell from the outside whether a lead had been seen, who had responded, or what had been said.
Chatwoot existed on the server but was being used only for finished patient records, not for live lead management. Evolution API was installed but not connected to any automation layer. n8n had two active workflows: one for appointment reminders (partially working) and one that had been paused for three months.
Week-by-Week Implementation Timeline
Day 1: TFCR Dropped to 3.8 Minutes Immediately
The first change was the simplest: a single n8n workflow connected to the clinic’s WhatsApp Business API instance via Evolution API that fired an automated acknowledgment message within 60 seconds of any new lead WhatsApp contact, regardless of time of day. The message was not generic, it was personalized to the lead source (Meta advertising leads received a version referencing “the ad you saw about dental treatment in Istanbul”) and included a qualification question: “So we can prepare the right information for you, can you let me know what treatment you’re looking into, implants, veneers, or something else?”
This single workflow change dropped the TFCR from 4 hours 17 minutes to 3.8 minutes on Day 1. The automated message is not a replacement for coordinator response, it’s a bridge. The patient knows they’ve been received. The coordinator has a 30–60 minute window to review the lead and personalize the follow-up. The urgency pressure is taken off the coordinator without removing their role.
Week 1: Chatwoot Routing and Lead Tagging
During Week 1, we configured Chatwoot to receive all incoming Evolution API conversations as tickets, with automatic assignment logic based on coordinator load. Each new lead was tagged by procedure type (from the qualification question response), by lead source (Meta advertising, Google Ads, WhatsApp organic, referral), and by time of contact. This gave coordinators a structured queue instead of a shared inbox with no context.
The coordinator team’s reaction was mixed. Two of the three coordinators adapted quickly, they found the structured queue easier to manage than the chaotic shared WhatsApp. The third coordinator initially ignored Chatwoot and continued managing leads from the personal WhatsApp number directly, creating duplicate records and missed handoffs.
Week 2: Supabase Integration and Lead Records
All incoming leads were now flowing into Supabase, name, phone number (normalized to international format), procedure interest, lead source, timestamp of first contact, and TFCR measurement. This gave us the first accurate data on lead volume, response time distribution, and coordinator attribution. The data confirmed the Day 1 TFCR improvement and revealed a secondary problem: 31% of leads who received the automated first response were not receiving a human follow-up within 4 hours. The automation had solved TFCR but exposed the next bottleneck.
Week 3: Coordinator Resistance to the New Handoff Format
The third coordinator escalated her resistance to Chatwoot during Week 3. Her specific objection: the structured Chatwoot format felt impersonal and she believed patients could tell they were talking to a system. This is a legitimate concern and worth addressing directly. The solution was not to force compliance but to show her the data: her leads handled through the personal WhatsApp number had a consultation booking rate of 9%. The leads handled through Chatwoot by the other two coordinators were booking at 19% (this was Week 3, the full system wasn’t yet in place). The data conversation was more effective than any process argument.
She began using Chatwoot in Week 4, with a modified message template that matched her personal tone more closely. By Week 6, her consultation booking rate was 28%.
Week 4–5: The Follow-Up Sequence for Non-Responders
Leads who received the automated acknowledgment but didn’t respond within 2 hours entered a separate n8n sequence: a follow-up message at 2 hours, a second follow-up at 24 hours (different framing, referencing the procedure they might be interested in), and a third at 72 hours. After 72 hours with no response, the lead was moved to a weekly nurture sequence in Supabase and removed from the active coordinator queue. This prevented coordinators from spending mental energy on cold leads while ensuring those leads weren’t permanently abandoned.
Week 6–7: Consultation Show Rate System
The pre-consultation 72-hour sequence (booking confirmation, 7-day check-in, 24-hour confirmation request) was deployed via n8n and Evolution API. Clinic D’s pre-intervention consultation show rate was 54%. By Week 8, it was 73%.
Week 8: Full Metrics Review
At eight weeks, all baseline metrics were measured against the same 30-day window in the prior period. The results are in the table above. The changes that produced the most measurable impact, in order: (1) automated TFCR fix on Day 1, (2) Chatwoot routing and coordinator queue structure, (3) follow-up sequence for non-responders, (4) consultation show rate sequence.
What Changed for the Coordinators?
The coordinators’ day-to-day experience shifted from reactive chaos to structured workflow. Instead of monitoring three separate WhatsApp threads while handling active patients, they worked from a Chatwoot queue with clear priority tagging, procedure context, and TFCR flags for overdue leads. Average coordinator messages per lead dropped from 22 to 9, not because coordinators were doing less, but because the automated acknowledgment and qualification message had already answered the first three questions every lead used to ask manually.
Coordinator capacity freed up by the efficiency gain was redirected to consultation-call quality, which is where the 21% to 34% deposit conversion improvement came from. The system didn’t replace coordinator work. It removed the coordination overhead so coordinators could do the work that actually converts.
What Is the Underlying Principle Here?
Lead Latency is the most expensive problem most clinics aren’t measuring. Clinic D was spending €4,200/month on Meta advertising to generate leads that were sitting in a shared WhatsApp inbox for four hours while patients chose competitors. The intervention cost less than one month of that ad spend to implement. The return was visible within 24 hours of Day 1. The full system took eight weeks to stabilize, including the coordinator resistance and retraining that is always part of any intake infrastructure deployment. The lesson is not that automation replaces coordinators. It’s that without automation in the TFCR layer, coordinator skill is irrelevant, because the patient has already moved on before a coordinator gets the chance to demonstrate it.
Frequently Asked Questions
What was the total cost of the intake system implementation for Clinic D?
The implementation involved n8n workflow configuration, Evolution API setup and connection to the clinic’s WhatsApp Business API instance, Chatwoot configuration and coordinator training, and Supabase schema build for lead and patient records. Total implementation cost was in the €6,000–9,000 range for a clinic of Clinic D’s profile. At the recovered revenue run rate of €18,000–26,000 per month, the system paid for itself within the first two to three weeks of full operation. Ongoing cost is server infrastructure plus the Evolution API subscription.
Why did Chatwoot matter more than just fixing the WhatsApp response time?
Because TFCR is only the first metric. A fast first response that flows into a disorganized shared inbox still loses leads, just later in the process. Chatwoot gave the coordinators a structured queue, gave management visibility into conversation status, and created the data foundation for measuring what happened after the first response. Without the Chatwoot layer, the Day 1 TFCR fix would have produced a temporary improvement that degraded back to the baseline as leads accumulated in the unmanaged system.
How do you handle lead responses that come in overnight or on weekends for a clinic that isn’t staffed 24 hours?
The automated acknowledgment via Evolution API fires regardless of business hours, patients at 11pm receive the same 60-second response as patients at 11am. The acknowledgment message is calibrated for out-of-hours contact: it confirms receipt, asks the qualification question, and sets a realistic expectation (“our team will follow up with more detail during business hours”). When coordinators arrive in the morning, the Chatwoot queue shows all overnight leads with their qualification responses already captured, sorted by arrival time, and flagged by procedure type. Out-of-hours leads that received the automated response and replied to the qualification question show a 28% higher consultation booking rate than leads that received no response until the next morning.
What happened to Clinic D’s Meta advertising CPL after the intake system went live?
CPL didn’t change immediately, the same ads, same targeting, same spend. What changed was the output per lead: more consultations from the same leads. By Week 8, the effective cost per consultation booked had dropped from €382 (€4,200/month ÷ 11 consultations) to €168 (€4,200/month ÷ 25 consultations). The ad spend became roughly 2.3x more efficient without any changes to the campaigns themselves. This is a common finding: clinics assume their Meta advertising is underperforming when the actual problem is intake latency killing the conversion rate downstream.
Is Clinic D’s result typical, or was their baseline unusually poor?
Their TFCR was worse than average but not extreme: I’ve audited Istanbul clinics with TFCRs above 8 hours. Their lead volume (80/month) and procedure mix (dental, mid-tier pricing) are typical for the Istanbul market segment. The magnitude of improvement (121% increase in consultation booking rate) was at the higher end, which reflects both how broken the baseline was and how much structural lead latency was suppressing an otherwise competent coordinator team. Clinics with a TFCR already below 30 minutes see smaller but still meaningful gains, typically 25–45% consultation rate improvement, when the full system (routing, follow-up sequences, show rate automation) is deployed.
Reviewed by Dr. Elif Sahin, Medical Director at MedTurkAI
*Running a clinic and not sure where your pipeline is leaking?*