Your AI outputs aren’t the issue. Your weak prompts are.

During Ragan’s AI Certificate Course, attendees learned that communicators need to reframe the way they interact with AI platforms. Dr. Alexandre Sévigny, associate professor of communications management at McMaster University, said that instead of loose prompts comms pros need to use more structured and repeatable prompt templates.

“The more structure you can build into your prompt, the better results you will get,” Sévigny said. If the direction you provide is fuzzy, the output will be generic.”

He added that AI prompt libraries are no longer optional—they’re essential to success with LLMs.

“Without them, you’re wasting time, creating inconsistency and even increasing energy usage with inefficient prompting,” Sévigny said.

Prompts also need to be tied to specific workflows to be most effective. For instance, there can be specific prompts for breaking down internal meetings into CEO memos, town halls into intranet posts and more. Sévigny told course attendees that specificity is a communicator’s friend when prompting.

“The more deterministic your prompts are and the more you define tone, structure, context and data, the better the output,” he said. “If you leave gaps, the model fills them with guesses.”