Personal voice

For AI tools to be used professionally, users need confidence that the output will be configurable and at least as consistent as what they would have produced by themselves. This could mean having the same tone of voice, for individuals and for brands. It could also mean including the same parameters or inputs into your prompts without having to write them over and over.

Tone of voice

Most generative writing platforms offer the ability to define one or more "voice" that represent a brand or individual. It would not be good if the content produced by Marketing team A sounded completely different than Team B.

Beyond the production efficiency, this has interesting use cases that could be explored in the future. For example, what might it look like to define your voice so your family can "talk" to you while you are away (perhaps, deployed or even in cases of death). How could we use this in research to understand how different personas might respond to a prompt.

This is closely related to shaping the AI Personality, though the AI's personality impacts how the AI interacts with you while the personal voice impacts how the AI reflects you back in its generations.

Key words and templates

Many tools are also allowing users and companies to store terms or information that can be drawn from in the future. For example, Copy.ai allows teams to save information about brands, products, trends, and so on and then introduce that context as a parameter into a prompt.

if a company had a specific way of referring to their product offering, this could be helpful when generating content to cover a launch.

Or, perhaps a user wants to set limitations about what should or shouldn't be contained in a response. ChatGPT allows users to set information that is included in prompts by default. For example, my bot is asked to tell me how it came up with an answer when prompted.

If the use cases behind your product include someone using the result for commercial or professional reasons, voice and tone controls will likely be critical. Think ahead of the use cases to find ways to differentiate. Could you change the voice of the chatbot, borrowing from Character.ai? Or perhaps you want to instruct the bot to provide responses that are more or less technical by default.

These are advanced parameters that will only make the technology more powerful, and more useful in commercial settings.

• Users may require the ability to store different styles of their voice for different uses
• The voice can be applied as a parameter, so the AI relies on the specific details of the voice when crafting its output
• Voices could be used for generating content for different channels, as in how a brand speaks to a LinkedIn audience vs. an instagram audience
• Consider how different individual and brand voices could be combined with patterns like a workflow to construct a personal message to any number of recipients in a common tone using the same references (like a release notes doc)