Case study · Saglo
A voice AI front desk for a shopping centre
Hundreds of tenants, one main phone line, and a front desk that couldn't keep up. We built Saglo a voice AI that answers every call in seconds, knows every store, and lands each caller in the right place — around the clock.
A phone tree is not a front desk.
A large shopping centre is a small city. Every day, the main line rings with store hours, directions, parking, gift cards, event questions, leasing inquiries, security issues, and callers asking for one specific store out of hundreds. The front desk did its best — and missed most of it. After hours, calls simply went nowhere.
The off-the-shelf answer is a phone tree: press 4, listen to options, guess. But a shopping centre's options change weekly — tenants open, close, and move — and nobody presses 7 with confidence. Callers gave up, tenants missed calls, and management had no idea what was being lost.
Saglo needed a front desk that never sleeps and never forgets a tenant. So we built one that speaks.
Every call answered in seconds. Every tenant reachable on one number. Every conversation logged, summarized, and followed up — by a voice that knows the whole centre by heart.
What we built.
Five systems, one voice — call receiving and routing for an entire shopping centre, end to end.
01
Every call answered, in seconds
No phone tree, no hold music, no voicemail black hole.
- A voice AI receptionist that picks up on the first ring, 24/7 — evenings, weekends, holidays, and the December crush when call volume triples.
- Natural conversation, not “press 1 for…”. Callers just say what they need — in English or French — and the AI keeps up with interruptions, accents, and half-finished questions.
- The AI answers the centre's common questions itself — hours, directions, parking, today's events, gift cards — so most calls never need a transfer at all.
02
It knows the whole centre
One brain trained on every tenant, every department, every service.
- A live tenant directory — stores, restaurants, services, kiosks — with hours, locations on the floor plan, and contact details, updated the moment a tenant opens, moves, or closes.
- Centre-wide knowledge baked in: parking rates and validate rules, transit access, accessibility services, lost and found, promo campaigns and in-mall events.
- When a tenant changes their hours for a holiday, the AI knows before the caller asks — one update, every call covered.
03
Routing that lands the call
The switchboard operator who knows everyone, working every line at once.
- Intent understanding that separates “where is the food court” from “I need to speak to the manager of the store that altered my suit” — and routes each one differently.
- Transfers to tenants, customer service, security, or the leasing office with a context whisper: staff hear who's calling and why before they pick up.
- Graceful handoff to human staff for complaints, sensitive matters, and anything the AI shouldn't handle alone — with a full summary attached, so callers never repeat themselves.
04
The work after the call
Every conversation becomes a record, a summary, and a follow-up.
- Automatic transcription and a short summary for every call — what was asked, what was answered, where it was routed.
- Message-taking that actually works: when a tenant can't take the call, the AI captures the message and delivers it to the right inbox with a transcript.
- Follow-up by SMS: callers get the store's hours, a map pin, or the event details they asked about, without writing anything down.
05
A dashboard for management and tenants
The phone system stops being a black box.
- Tenant-level views: each store sees its own calls, messages, and missed opportunities; centre management sees the whole picture.
- Pattern analytics — what callers ask for, when volume spikes, which questions repeat — that feed staffing, signage, and event planning decisions.
- An escalation audit trail: every call handed to a human is logged with its reason, so nothing falls through the cracks.
Need a front desk that never sleeps?
Tell us what your calls need to do. We'll tell you honestly whether we're the right team to build it.