What was the call-centre volume deflected for 123 Dentiste?

32.5% — Call-centre volume deflected at 123 Dentiste using Vatel AI voice agents.

What was the calls triaged per week for 123 Dentiste?

1,600 — Calls triaged per week at 123 Dentiste using Vatel AI voice agents.

What was the weekly talk time on ai for 123 Dentiste?

590 min — Weekly talk time on AI at 123 Dentiste using Vatel AI voice agents.

All case studies

HealthcareDental networkQuebec, Canada

123 Dentiste

A Quebec dental network deflects a third of its call centre volume with an AI front desk.

123 Dentiste added Vatel as an AI triage layer in front of its central call centre, deflecting routine requests, differentiating new vs. existing patients, and routing calls across clinics 24/7.

EngagementOngoing · inbound triage

RegionQuebec, Canada

ScopeInbound triage · 24/7 coverage

LanguagesFrench + English

StatusLive in production

Locations40+ clinics

Call volume~1,600 calls / week

32.5%

Call-centre volume deflected

1,600

Calls triaged per week

590 min

Weekly talk time on AI

01The challenge

A call centre drowning in routine volume.

123 Dentiste operates one of the largest dental networks in the country, with over 40 clinics served by a shared central call centre. As the network grew, call volume outpaced headcount, and routine requests were eating up agent time that should have gone to complex cases.

Routine calls dominated the queue. Around a third of calls were simple inquiries, hours, directions, insurance acceptance, appointment availability, that required no agent judgment but still consumed agent time.
No after-hours coverage. Calls outside business hours went to voicemail and were rarely returned promptly. Patients with time-sensitive needs had no way to reach the network outside of a narrow window.
New vs. existing patients indistinguishable at intake. Agents had no way to identify caller type before engaging, leading to misroutes and repeated information gathering across the network.

02The approach

A triage layer that acts before the call centre answers.

We designed the agent as a front layer, not a replacement, for the existing call centre. It needed to classify and resolve what it could, and route everything else with context attached.

01Wk 1

Discovery

Call type audit
Routing map
CRM integration review

02Wks 2–3

Build

FR + EN call flows
New vs. existing patient logic
CRM lookup integration

03Wk 4

Deployment

SIP configuration
Live monitoring
Agent briefing

03The solution

Every call answered immediately, routed or resolved in seconds.

The agent answers every inbound call, classifies intent, differentiates new from existing patients, and either resolves the inquiry or routes to the right agent with a context summary.

Intent classification at pickup

The agent identifies on the first exchange whether the caller needs information, an appointment, a transfer, or an escalation, and acts accordingly without menu navigation.

New vs. existing patient differentiation

Callers are identified as new or existing patients early in the call, routing them to the appropriate agent or flow without requiring the centre to make that judgment.

24/7 inbound coverage

The agent runs around the clock. After-hours calls are resolved where possible and logged with full context for the morning queue.

FR/EN bilingual operation

The agent detects language on the first exchange and conducts the full conversation in French or English without a selection prompt.

04Results