AI Horizons 2026: What PR and Comms Pros Need to Do Next
AI is no longer a side project. It's a daily part of the job - from content and internal comms to stakeholder intel and measurement.
At Ragan's AI Horizons Conference in Fort Lauderdale, speakers cut through hype and talked execution. Here are the moves worth making now.
Commit to change before chasing tools
Charlene Li made it clear: most teams don't stall because AI "doesn't work." They stall because leaders won't commit, don't communicate the plan, and avoid guiding people through change. Strategy first, tech second.
Michelle Bauer stressed psychological safety. If people don't feel safe experimenting (and failing), adoption slows no matter how good the platform is. Build trust, then build capability.
Christine de Valle showed how AI fluency grows through shared experimentation. Create spaces to swap use cases, demos and quick wins. Confidence compounds when peers teach peers.
Brenden Lee gave people an entire day to play with AI. Time is the unlock - not another tool. Put experimentation on the calendar.
Partner with IT and fix your workflows
Martin Waxman urged communicators to build real relationships with IT. Know what data, systems and permissions you already have - they're often better than you think, and IT can help you scale safely.
Alex Sévigny was blunt: "Train processes and workflows, not prompting." Your gains come from standardizing how work moves, not from clever one-off prompts.
Chris Gee reminded us that AI agents replace tasks, not whole jobs. Offload repetitive steps so your team can focus on delicate calls, nuance, and stakeholder judgment.
- Map one high-volume workflow (e.g., media monitoring to exec brief) and standardize it with AI.
- Document inputs, approvals, and outputs so it's repeatable and safe.
- Meet monthly with IT to expand access, automate steps, and review risk.
Internal comms: the 80/20 that actually ships
Sarah Whitty shared a simple rule: let AI get you to 80%. Then add strategy, tone, and the political read before anything goes out. That last 20% protects trust.
Michael A. Cousin showed why personas matter when your workforce ranges from bus drivers to nuclear scientists. Different roles need different context, channels, and reading levels - train your AI to reflect that.
- Create 4-6 employee personas and tag messages by persona, channel, and reading level.
- Use AI for drafting, translation, and variants; humans finalize intent, nuance, and timing.
- Measure opens and questions by persona to refine guidance and templates.
AI video is now a comms skill
Stephanie Nivinskus suggested prompting like a film director: shot list, pacing, visual style, voiceover beats, brand guardrails. Better inputs mean safer, on-brand outputs.
Steve Van Dinter said the tools are ready for business use. Start small pilots and build a reel of approved formats (explainers, FAQs, CEO notes, policy updates).
- Build a "director prompt" template with scenes, brand rules, and approvals.
- Caption everything, test mobile-first, and store reusable b-roll and VO scripts.
GEO, Reddit, and the long game
Jenna Galbreath pointed out that Reddit matters for GEO - and it punishes promotion. Be helpful, or you'll get banned. Start seeding useful content now so you have credibility later.
Will Hodges called GEO a return to the basics of PR: clarity, relevance, proof, repetition. Help people solve real problems and let communities amplify it.
- Contribute answers, case details, and tools users actually need. Avoid brand chest-thumping.
- Review Reddit's self-promotion rules before you post. Read the guidelines.
Guardrails that prevent headlines you don't want
Marcel Williams advised checking both sides of the system: inputs used to train and the outputs you publish. Without scrutiny on both, bias incidents escalate fast.
Rowan Toffoli reminded everyone that AI isn't just about speed. Use it to improve analysis, creativity, testing, and clarity - not just to do the same work faster.
A simple playbook for the next 90 days
- Declare a leadership-backed AI plan with 2-3 high-impact use cases.
- Schedule hands-on time for teams to test tools and share wins.
- Standardize one workflow end to end; document it; train it; measure it.
- Meet with IT and Legal monthly to expand safe capabilities.
- Adopt the 80/20 rule for internal comms and enforce persona tagging.
- Launch one AI video format and build a reusable "director prompt."
- Join 2-3 relevant subreddits; contribute helpful content weekly, no promotion.
- Audit for bias at input and output; keep a log of fixes and decisions.
Want structured upskilling?
If you need a clear path to build skills by role, explore this collection of AI courses by job: Complete AI Training: Courses by Job.
Your membership also unlocks:
