Hi, and Happy Friday!

At our Q1 AI Roundtable webinar, Beyond ChatGPT: The AI Toolkit Every Communicator Needs in 2026, we got into the nitty-gritty of integrating AI into comms workflows. The discussion spanned research, drafting, stakeholder intelligence, and content production.

T2B Pro members can click here to watch the replay and browse a comprehensive list of the resources shared.

Top 10 Takeaways

1. Recognize that AI can start to lead you. Front-loading context and expertise keeps you in the driver's seat rather than reacting to what the model produces.
2. Treat your comms expertise as training data. Decades of audience knowledge and stakeholder nuance can't be inferred by AI; getting it into structured training documents separates generic output from work that reflects real professional judgment.
3. Use voice-to-text to surface knowledge on autopilot. Speaking prompts tend to surface more nuance and professional instinct than typing, reinforcing what the model prioritizes in its output. Plus, it’s much faster than typing.
4. Prime AI before you prompt it. Load brand voice, technical specs, and context before starting any task, especially across multiple brands or regulatory environments.
5. Build a prompt library, not just a prompt habit. Reusable brand voice guides and context documents function like institutional knowledge, allowing even junior team members to produce senior-level outputs. (Pro tip: Link the library to your LLM of choice.)
6. Distinguish between AI as brainstorm partner and AI as output engine. Conversational chat for thinking through an approach is a different workflow than structured prompting toward a finished deliverable, and conflating the two tends to produce results that feel neither strategic nor polished.
7. Understand that not all models are built for the same work. The distinction between higher-reasoning and more creatively flexible models matters, and within a single platform, the model tier selected can dramatically change output quality.
8. Consider AI-powered tools for stakeholder landscape research. Deep research modes can map how HCPs, advocacy organizations, and patients discuss the same topic in real time, giving comms teams a faster baseline for message development.
9. Know that enterprise accounts protect confidential work. Business-level subscriptions segment conversations on separate infrastructure and don't use prompts for model training, which matters when working with pipeline assets or pre-announcement materials.
10. Recognize the window of subsidized experimentation. Current AI pricing doesn't reflect true model costs and is expected to rise, making now an unusually low-friction time to build workflows before price becomes a barrier.

Remember, AI is most useful when it amplifies existing communications judgment, and like any tool, it is only as good as the person using it.

Huge thanks to Nathaniel Reagan (AI Learning Labs) and Matt Snodgrass (Syneos Health Communications) of the T2B AI Roundtable for leading another great session -- I'm still getting notes about it a couple weeks later! And also thank you to everyone who joined the discussion and shared the tools and workflows they’re experimenting with.

Lynnea