3wordTaxi App

The project tested whether voice commands could provide a compelling and reliable way to request transportation, while demonstrating the commercial potential of integrating what3words with established ride-hailing services. Research and testing focused on voice-recognition accuracy, user trust, conversational interaction and seamless third-party API integration.

The released concept received positive feedback from users, investors and partners for making taxi booking and destination entry more convenient.

3wordTaxi was a voice-enabled taxi booking concept built around what3words, which resolves every 3m x 3m square in the world to a unique three-word address. The app let users speak a destination, see the recognised phrase, confirm the destination and ETA, and then hand off the request to Uber or another ride provider. It was designed particularly for driving contexts, where users need precise navigation without taking their eyes off the road or hands off the wheel.

The central challenge was proving that a voice-enabled taxi booking experience could be robust and reliable enough for users to request a ride using only their voice. The app also needed to feel conversational rather than robotic, support a hands-off interaction, solve the ambiguity and pronunciation problems of conventional addresses, and maintain a consistent experience across multiple taxi providers and their APIs.

Traditional postal addresses perform poorly with voice input because similar-sounding streets and numbers are easy to confuse, unfamiliar names can be difficult to pronounce, and house or road names are not always unique. These problems are especially risky when directing a vehicle to a precise entrance or pickup point. what3words created an opportunity to replace lengthy, ambiguous addresses with memorable three-word locations that are precise and easier to speak.

The product needed to let a driver state a precise destination without typing an address, searching a map or manually correcting a dropped pin. It also had to make recognition errors visible and recoverable, ensuring that a misheard three-word address would not immediately result in an incorrect taxi booking.

The objective was to test and demonstrate an end-to-end voice taxi journey: invoke the experience, speak a what3words destination, receive clear visual, audible and haptic feedback, confirm the recognised destination and ETA, and complete the booking through an integrated ride-hailing provider.

The single headline goal was to prove that users could request a taxi and navigate to a precise three-word address through a reliable, hands-free voice interaction.

The project followed a general UX and Product Design workflow from scope and requirements through research and analysis, concept and UX design, production design, development, sign-off and iteration. The team used continuous-improvement loops to feed findings back into the design. The approach was informed by the growing adoption of Amazon Alexa, smartphone voice queries and other voice-controlled products, while explicitly addressing early voice technology concerns around accuracy and trust.

Research examined how to make the voice interface feel like everyday conversation, including the cultural expectations and implied meanings that sit beyond literal words. Early exploration tested voice-wave visual designs, sounds and haptic feedback. Sketches covered the required input and output text, listening states, calls to action into Uber, layered waveform behaviour and multisensory progress feedback for drivers who could not continuously look at the screen.

Extensive user testing informed changes to functionality and ease of use.

The solution combined what3words precise addressing with a conversational taxi-booking flow. A large "Where would you like to go?" prompt initiated the interaction; the spoken destination appeared as live text, was read back for confirmation, and was paired with an estimated arrival time before the request proceeded. After confirmation, the experience handed the destination into Uber's map and ride-selection flow, while the what3words system could also be integrated into automotive infotainment systems.

The V1.0 task model mapped a voice-plus-hands Uber booking journey from Siri-style invocation into a loading state, destination prompt, spoken three-word address capture, confirmation branching and request processing. It documented which feedback channel applied at each stage, including voice and sound, sound only, waveform and text, or haptic feedback.

The wider Visual Conversation flow separated pickup and destination, retained the location summary, and defined the handoff into the standard Uber request screen.

The voice-wave concept tested eight tiled animation studies surrounding the fixed what3words mark. Brightness and contrast changes in the grid communicated whether the system was listening or speaking without depending on a conventional microphone icon or explanatory copy. The final interaction used a large voice prompt, live recognised text, a distinct processing state and a persistent pickup/destination summary to keep the experience calm and glanceable in an automotive context.

The research established that spoken destinations should be transcribed and read back for explicit validation before a booking became consequential. A "no" response needed to loop back into correction rather than ending the journey. The flow also introduced provider comparison, for example stating that an Uber could arrive in three minutes while a Lyft could arrive in five, before asking which provider to book.

The Voice Taxi app received positive user feedback for offering a unique and convenient way to request a taxi without typing or opening a separate app. Its reception among potential investors and partners supported the viability of voice-enabled transportation and IoT products. The work contributed to what3words integrations in Mercedes, Land Rover and TomTom products, as well as compatibility with Amazon Alexa and Google Home, demonstrating the value of precise voice-friendly addressing.

Voice interfaces need to make recognition visible and consequential actions reversible. In this project, precise three-word addressing solved the ambiguity of conventional spoken addresses, while read-back confirmation, correction paths and multisensory feedback built trust for a hands-free interaction.

A later concept extended the confirmation experience beyond the car to a smartwatch. Shown over an aerial view of Tower Bridge, it combined the question "Would you like to confirm the order?" with the car ETA, what3words mark and pickup address, suggesting a future multimodal system in which users could confirm a hands-free booking from an appropriate wearable touchpoint.

An interactive prototype tested the conversational design from voice input through booking confirmation. It displayed the current pickup as a three-word address, read back a resolved destination such as "index.home.raft," stated an ETA such as four minutes, and asked "Would you like to confirm the order?" This made dispatch an explicit and reversible decision rather than an immediate consequence of speech recognition.

Because drivers could not safely rely on continuous visual attention, the interface used visible, audible and tactile feedback together. Live transcription showed what the system had heard, while the tiled waveform indicated the listening state and sound and vibration communicated progress. A separate dark "...processing request..." state clarified when the system was working, helping users distinguish between moments when they should speak and moments when the request was being processed.

The concept was validated through voice-flow wireframes, visual feedback explorations, an interactive prototype and extensive user testing before release. Confirmed bookings were handed into Uber's own map and ride-selection interface, while the broader concept was positioned for integration with other providers through their APIs. Related what3words integrations subsequently extended into automotive products and smart-assistant environments.

Kindly be advised that this portfolio has been carefully curated to comply with all applicable company policies and regulations. Any omissions of information or details are deliberate and in strict accordance with confidentiality agreements and legal obligations. Your understanding of this compliance measure is greatly appreciated.

The material presented here is for portfolio purposes and follows the Fair Dealing guidelines. No confidential or proprietary information from what3words Ltd has been disclosed.

This project is a representation of my work and does not necessarily reflect the views or strategies of the what3words Ltd.