Conversational AI & voice design

I worked with Digitas on an early conversational AI and voice experience for Kellogg’s Special K, researching, developing and writing several months’ worth of content for Amazon Alexa.

At the time, voice assistants were at the forefront of consumer technology, and there were few established conventions for how brands should communicate through AI.

A key part of the project was therefore shaping the AI voice itself: defining how Special K should sound in conversation, how information should be structured for a voice-first experience, and how the brand’s personality could translate from traditional channels into a responsive digital assistant.

The resulting Healthy Living Tips Alexa experience combined conversational writing, UX thinking and tone of voice to deliver useful content in a format that felt natural when spoken rather than read.

Achievements

• Helped shape the voice and personality of an early brand-led AI experience, defining how Special K should communicate through Alexa.
• Developed conversational principles and tone of voice for a voice-first interface, translating an established consumer brand into a new interactive medium.
• Researched and wrote several months of Alexa scripts, creating a substantial library of conversational content for the Healthy Living Tips experience.
• Designed content specifically for spoken interaction, considering brevity, clarity, natural language, pacing and how users process information without a visual interface.
• Applied emerging conversational UX principles at a time when voice technology was still new, helping establish patterns for how the experience should respond and communicate.
• Turned health and lifestyle information into concise, accessible spoken content, balancing usefulness with an engaging and recognisable brand voice.
• Maintained consistency across a large volume of scripted interactions, ensuring the AI experience felt coherent rather than like a collection of individual pieces of copy.
• Worked at the intersection of creative, content and emerging technology, gaining early experience of a challenge that has since become central to generative AI: how to give an AI interface a distinctive, consistent and appropriate voice.