Intro to a communicator's approach to custom AI bots

The big idea

Building a custom AI bot is a communications strategy exercise, not a technical one. The same skills you use to create a brand voice guide—understanding audience, voice, style, and channels—are exactly what you need to train a bot. ​“One of the best groups of people to create bots is professional communicators,” says Samantha Becker.

The framework

The core insight: ​“Building a bot is no different from creating a communication strategy. You need to know your organization’s voice, style, audience, and channels. All AI is doing is dipping into your knowledge base.”

When to use this

• Your organization has multiple departments or brands producing content with inconsistent voice
• You have limited communications staff but need to scale storytelling
• Non-communicators (program directors, faculty, executives) need to create on-brand content
• You want to onboard new team members faster on organizational voice

Common mistakes

• Skipping the strategy work: jumping straight to the AI tool without first documenting voice, style, and brand positions results in a bot that produces generic output
• Building once and forgetting: a bot trained once degrades over time; schedule regular updates, gather user feedback, and monitor for drift
• Over-restricting access: the value multiplies when everyone in the organization can use the bot, not only the communications team
• Expecting perfection: AI outputs always require human editing; use the bot for first drafts and to overcome blank-page syndrome, not as a replacement for editorial judgment

Quick start

Start by consolidating your existing brand voice documents into a single file. Upload it to your AI platform and create a custom bot with clear instructions about its role. Test it against generic ChatGPT on the same prompt. The difference will show you what’s possible.