Flagship product

AnyTaxi

AI Voice Fleet Operations Platform

Starting with the phone call, AnyTaxi converts inbound orders into traceable, structured records and connects them to dispatch and fleet operations.

Discuss a fleet pilot View development status
  • Voice order intake
  • Address extraction
  • Dispatch integration

Positioning

Not just transcription — turning calls into executable orders.

AnyTaxi is designed to preserve the phone-based ordering habits fleets already have, while producing structured and traceable operations data. That requires voice processing, order workflows, and dispatch systems to be designed together.

Keep the phone channel

Passengers keep calling as usual, reducing behaviour change at adoption.

Turn speech into fields

Extract pickup, time, and passenger details, while flagging information that needs human confirmation.

Make operations reviewable

Orders, processing states, and review records form traceable operations data.

Order workflow

Five inspectable stages from call to dispatch.

This is the current workflow design. Each stage includes failure handling and human intervention points so ambiguous information does not flow directly into dispatch.

  1. 01

    Inbound call

    A voice workflow guides the caller to provide pickup, time, and trip details.

  2. 02

    Speech-to-text

    Call audio becomes text, creating a traceable processing record.

  3. 03

    Field extraction

    Address, time, passenger count, and notes are extracted into structured fields.

  4. 04

    Human review

    Low-confidence or ambiguous details are confirmed by a person before proceeding.

  5. 05

    Order & dispatch

    The order is created and connected to dispatch, producing traceable operations data.

The workflow reflects the current design direction. Actual capabilities depend on development and testing progress.

Development status

Capability status is stated plainly. Planned work is not described as live.

The table reflects current development progress. AnyTaxi is still in R&D and validation. It is not a publicly available commercial service, and no recognition accuracy or customer outcome figures are published.

  • Order workflow and data field definitionCompleted

    Workflow scripts, required fields, and exception scenarios have been defined as a shared baseline for development and testing.

    Completed
  • Speech-to-text and address extractionIn validation

    Recognition and field extraction are being tested with call samples, including adjustments for spoken address variations.

    In validation
  • Order creation and dispatch integrationIn development

    The order data model and integration interface are being built. Integration testing continues; no public service is live yet.

    In development
  • Fleet operations data viewIn development

    Designing how calls, orders, and processing states are aggregated so operations teams can review outcomes.

    In development
  • Multilingual voice workflowPlanned

    Language and scenario priorities will be assessed with pilot requirements. Implementation has not started.

    Planned
  • Automated dispatch decisionsPlanned

    Requires sufficient operations data and rule validation before automation scope and human review can be assessed.

    Planned

Concept architecture

Layered design so voice, orders, and dispatch can be validated separately.

Voice processing and the operations core are separated by clear interfaces, so human review, testing, and future replacement do not require rewriting the whole system. Deployment details and sensitive information are not disclosed publicly.

  • Call audio pre-processing and speech-to-text
  • Spoken address variants and field normalisation
  • Human review for low-confidence information
  • Decoupled order service and dispatch interface
  • Access control and data minimisation
Concept architecture. Deployment details and sensitive information are intentionally omitted.

Pilot cooperation

Start with a well-defined pilot and validate together.

We are currently focused on cooperation discussions and pilot planning; commercial pricing is not published. The best fit is an operations team with an existing phone order workflow, willing to define success metrics and provide de-identified samples and feedback.

Email about a pilot

Partner names, logos, and customer data are not published without written permission.

  1. 01

    Define pilot scope

    Agree on order scenarios, participants, data boundaries, and success metrics.

  2. 02

    Sample testing

    Use de-identified call samples to check the voice and field extraction workflow.

  3. 03

    Set up review

    Define how low-confidence information is handled and where humans intervene.

  4. 04

    Iterate from feedback

    Adjust workflows, fields, and integrations based on real usage feedback.

Work with us

Turn phone orders into traceable operations data.

Fleet operators and enterprise partners are welcome to discuss a pilot or request publicly shareable product and R&D information.

Business email Contact us