Une société de gestion immobilière fait fonctionner toute son réception téléphonique par l'IA.
Devpro built a virtual front-desk agent for Golden Equity, a property management company operating across Ontario and Quebec. The agent replaced the receptionist function entirely, handling leasing inquiries, operational calls, and administrative routing for roughly 67 inbound calls per day.
- Engagement
- 4 weeks · signed Feb 17 · live Mar 18
- Region
- ON & QC, Canada
- Scope
- Full front-desk call handling
- Volume
- ~67 calls / day
- Status
- Live in production
Sixty-seven calls a day through a single receptionist.
Golden Equity's inbound call volume had grown past what a single front-desk contact could reliably manage. Calls arrived across leasing, operations, and accounts, each requiring a different contact, and the receptionist had no way to handle peak volume, after-hours calls, or the routing complexity that came with a multi-building portfolio.
- Routing required building context. Leasing inquiries couldn't be forwarded to the corporate agency without first identifying which property the caller was asking about. That qualification step required a person.
- Operations calls needed the right specialist. Sending a maintenance call to the wrong building manager or a general line created delays and callbacks. Callers expected to reach the right person on the first try.
- Admin calls were drowning the front desk. Accounts payable and receivable calls were mixed in with leasing and operational volume, leaving the receptionist unable to prioritise or triage effectively.
A virtual agent that qualifies, routes, and never misses a call.
We designed the agent around Golden Equity's actual call taxonomy, mapping the three distinct caller intents and the routing logic behind each before writing a single conversation flow.
Discovery
Audited inbound call history, catalogued call types, mapped routing destinations, and documented the qualifying questions the receptionist was already asking.
- Call type audit
- Routing destination map
- Qualifier extraction
Build
Built the three core conversation flows: leasing inquiry with building identification, operational issue triage with specialist routing, and accounts routing. Configured transfer logic for each destination.
- Leasing intent flow
- Operations routing flow
- Accounts routing flow
Testing
Validated all routing paths across simulated call scenarios. Tested edge cases including callers who couldn't identify their building and calls that crossed multiple intent categories.
- Routing path validation
- Edge case testing
- Transfer accuracy review
Go-live
Deployed to production as the primary front-desk contact number. Post-launch monitoring through the first weeks of live volume.
- Production cutover
- Live monitoring
- Post-launch adjustments
One agent. Every call. The right contact every time.
The virtual agent greets every caller, identifies the purpose of the call, asks the qualifying questions specific to each intent, and routes to the appropriate contact without involving the receptionist.
Intent classification
The agent identifies on the first exchange whether the caller has a leasing inquiry, an operational issue, or an administrative matter, and routes the conversation accordingly without requiring menu navigation.
Building identification for leasing
Leasing callers are asked which property they are inquiring about before the transfer. The corporate leasing agency receives the call with the building context already established.
Operations specialist routing
Operational issues are triaged by type and routed to the appropriate building manager or operations specialist, eliminating the callback loop that came from generic forwarding.
Accounts routing
Accounts payable and receivable callers are identified and routed to the correct contact without going through the front desk, freeing the receptionist from a category of call they couldn't resolve.
24/7 call coverage
Calls outside business hours are handled by the agent, with routing to voicemail or scheduled callbacks depending on the call type, ensuring no inquiry is lost to after-hours cutoff.
Sixty-seven calls a day handled without a receptionist.
Measured since deployment, against the prior state where all inbound volume passed through a single in-house contact.
Calls handled by AI daily
Full front-desk volume without receptionist involvement
Routing destinations served
Leasing agency, operations specialist, accounts
Call coverage
After-hours calls handled and logged automatically
Calls lost to peak overflow
No missed calls due to single-contact capacity
