AI and Automation: Communicators are lagging on these AI skills. Here's how to catch up.
Most comms teams aren't behind on tools. They're behind on the foundations AI needs to work. As Alex Sevigny, associate professor of communications at McMaster University and longtime advisor to PR leaders, put it: "You can't use AI in communications until you can clearly explain how you do your job."
He's right. Many of us can describe outcomes, but not the steps. And without clear steps, data access, or cross-functional partners, AI becomes guesswork.
3 gaps your team should zero in on now
1) Build strong allies
Communicators often operate in a bubble. "We don't necessarily have friends or colleagues we talk to constantly in IT, data science or finance," Sevigny said. That isolation keeps you out of the rooms where systems, data and automation decisions are made.
Start with people, not platforms. Bring IT, data, and finance into your world and ask them into yours. "Tell them, 'This is what I do hour to hour. This is how I produce my work.' Ask how to make it more 'AI friendly,' what data could help, and what questions you should be asking."
- Set three 30-minute coffees: IT, data, finance. Share your hour-by-hour process, not your job title.
- Ask: What data already exists that could support this? What approvals or risks should I plan for? What's the fastest safe experiment we can run?
- Co-create a tiny pilot (one workflow, one audience, one measurable outcome).
2) Define your work
"If you can't tell me blow-by-blow how you produce a writing product, you can't really automate it in a safe or credible fashion," Sevigny said. Other professions have detailed specs and checklists. PR often has tribal knowledge and Slack threads.
Pick one high-volume task and map it precisely. The goal isn't bureaucracy. It's clarity, repeatability, and safer automation.
- Choose a core task (press release, media pitch, executive memo, investor deck).
- Write the step-by-step: inputs, sources, decisions, reviewers, approvals, tools, timing, and criteria for "done."
- Mark risks: legal, brand, privacy, disclosure, accessibility, and any must-check facts.
- Turn it into a checklist and prompt template so AI can assist without guessing.
Once the workflow is explicit, you can test AI at specific steps: research synthesis, outline drafts, version control, or formatting. No mess, less risk.
3) Understand how to use data
Many communicators come from journalism or humanities. Numbers weren't the focus. But AI runs on data. "If you're going to use AI effectively in a way that's verifiable and defensible, you have to understand what data it's working on," Sevigny said.
Your organization already collects a lot: email performance, CMS analytics, CRM records, stakeholder notes, survey data, social listening, support tickets. If you don't know what exists, who owns it, or what's off-limits, you can't validate outputs or prove ROI.
- Inventory systems you touch weekly (email, CMS, CRM, social, events, intranet, help desk).
- For each: what data is captured, where it lives, retention rules, who owns it, and what you're allowed to use.
- Define "truth" sources (metrics that matter to leadership): reputation indicators, engagement by audience, message recall, and business tie-ins (pipeline, retention, cost avoidance).
- Pilot a simple validation loop: AI output → human review → data check → revision → final. Save examples to improve prompts and policies.
You don't need to be a data scientist. You do need data literacy and guardrails. Courses help, but the fastest gains come from sitting down with internal owners and asking clear questions.
Suggested next steps and resources
- Skill up with peers: Ragan Training offers courses on workflows and data-driven comms.
- Build a practical learning path by role: AI courses by job for communications pros.
30-60-90 day plan
- Days 1-30: Meet allies in IT/data/finance. Document one core workflow end-to-end. Inventory data sources and permissions.
- Days 31-60: Run a small pilot on that workflow (e.g., AI-assisted outlines or message variants). Create a prompt library and checklist. Collect feedback and examples.
- Days 61-90: Add measurement (before/after time saved, quality scores, stakeholder satisfaction). Present results to leadership. Publish lightweight guidelines for AI use.
Bottom line
AI doesn't start with prompts. It starts with people, process, and data. Build allies, define your work, and know your data. Do that, and AI becomes practical, safe, and genuinely useful for PR and communications.
Want more? Attend Ragan's AI Horizons Conference, Feb. 2-4 in Ft. Lauderdale, FL, to learn from industry experts and see how others are putting this into practice.
Your membership also unlocks:
