Inside the AI receptionist

See exactly how the AI receptionist handles a call.

Follow how Leo identifies the request, uses approved business rules, completes an agreed action, and uses the handoff or follow-up path when a person is needed.

Your number stays yours. The call flow is configured before launch.

Sample callA sample call, switching languages mid-sentence:
Hi, do you have anything Thursday afternoon?
9:42
Let me check… Thursday at 2:30 or 4:00 — which works better?
9:42
Las 4 está bien. ¿Podemos seguir en español?
9:42
¡Claro! Le agendo el jueves a las 4. ¿Me da su nombre, por favor?
9:43
María López.
9:43
Gracias, María — la confirmación por SMS ya está en camino. ¡Hasta el jueves!
9:43
Booked into the calendarSMS confirmation sent
Your existing number stays in placeAnswers use approved business informationEach deployment language is testedA message or handoff path remains the fallback
One call, end to end

What happens after the phone rings

The assistant does not just answer. It moves the call through an approved path and leaves a result someone can act on.

  1. Stage 1

    The call is forwarded

    Your existing number sends the calls you choose—such as after-hours or unanswered calls—to Leo.

  2. Stage 2

    Language and intent are identified

    The assistant identifies itself as a virtual assistant, then establishes the caller's language and request.

  3. Stage 3

    The approved workflow runs

    It uses only the business information and rules it is allowed to use, then books, captures, or routes the request.

  4. Stage 4

    The outcome is delivered

    The appointment, message, transfer, or follow-up reaches the destination agreed for that call type.

Call handling by scenario

Different requests, different rules

Each call type gets its own approved path. The assistant does not improvise one generic answer.

New appointment

Caller request
“Do you have anything Thursday afternoon?”
Assistant action
Checks approved availability and collects the details required by your booking rules.
Business outcome
Books the slot or records a precise follow-up request if the booking connection is unavailable.

Existing appointment

Caller request
“I need to move or cancel my appointment.”
Assistant action
Uses the approved rescheduling path and confirms the information needed to make the change.
Business outcome
Updates the agreed workflow or sends a change request to the responsible person.

Service or quote request

Caller request
“Can someone price this job?”
Assistant action
Collects the service, location, timing, and contact details you define—without inventing a price.
Business outcome
Delivers a qualified message with the information your team needs to respond.

Urgent or uncertain request

Caller request
“This is urgent; I need to speak with someone.”
Assistant action
Checks the escalation conditions and uses the transfer or follow-up path approved for that situation.
Business outcome
A person receives the call or a priority follow-up with the relevant context.
Phone line and destinations

Fits around the number and tools you already use

Leo is added through call forwarding. The routes below are selected during onboarding; no replacement number is required.

Your existing business number

Forward only the call types and time windows you want Leo to handle.

Your configured Leo workflow

The approved greeting, business details, call paths, deployment languages, and fallback rules are applied during the call.

Where the result goes

Calendar or booking flow

Completed bookings use the agreed calendar workflow. If it is unavailable, the request becomes a follow-up.

Inbox or message workflow

Caller details, the request, and the next step are sent in the concise format agreed during setup.

A person on your team

Urgent, uncertain, or explicitly requested conversations use the configured transfer or follow-up path.

Every destination and fallback is confirmed before launch. Custom integrations are scoped separately.

Multilingual deployment

Language support is configured, not assumed

The underlying voice engine supports 90+ languages. A production setup includes only the languages selected and validated for your business.

English and Turkish have already been validated end to end. Every additional deployment language is tested against the business's real call scenarios before launch.

See the engine language list
  1. 1

    Select real caller languages

    Choose the languages your customers actually use rather than enabling an untested list.

  2. 2

    Review business terminology

    Services, names, policies, dates, and handoff phrases are checked in each deployment language.

  3. 3

    Test realistic calls

    Bookings, questions, mid-call language changes, and fallback scenarios are reviewed before launch.

Operating boundaries

Clear rules for when the assistant should stop

A reliable receptionist is defined as much by its handoff rules as by what it can complete.

Handled inside the approved flow

  • Book, move, or cancel appointments using approved availability
  • Answer from reviewed business information
  • Take messages and collect the agreed contact details
  • Use configured confirmation and transfer workflows

A message or transfer takes over when

  • The requested information has not been approved
  • A calendar or required integration is unavailable
  • The request needs judgment outside the configured workflow
  • The caller asks for a person
Configuration contract

A production-ready setup starts with your real call rules

Before launch, we document the information a receptionist would normally learn from you and the limits it must follow.

You provide

  • Services, hours, locations, policies, and approved answers
  • Appointment types, availability, change, and cancellation rules
  • Who receives transfers, what counts as urgent, and where messages go
  • Deployment languages and representative call scenarios

We configure and validate

  • Virtual-assistant greeting and caller disclosure
  • Approved call paths, data fields, and outcome formats
  • Integration and fallback behavior for each scenario
  • Pre-launch test cases and review checkpoints

You receive the documented call flow and onboarding scope before the system goes live.

Bring one real call scenario.

We will map how Leo should answer it, what it may complete, and when it should transfer or create a follow-up.