If you've used ChatGPT or Claude for press releases, pitch emails, or media briefings, you might notice that while some outputs are decent, many tend to be overly generic. The issue isn't the AI itself, but rather how you're prompting it.
Many PR pros miss out on significant value because they treat AI like a search engine rather than a collaborative partner. The good news is that learning effective prompting isn't difficult; it simply involves grasping a few essential principles that can turn average results into highly useful initial drafts.
Why Most PR Prompts Fail
A typical prompt might be: "Write a press release about our new product launch." The AI responds with a generic piece that sounds like every other press release. It's technically accurate but utterly unmemorable, which is the last thing you want when trying to capture a journalist's attention.
The problem isn't the AI's capability — it's that you've given it almost no information to work with. The AI doesn't know your brand voice, your target audience, what makes this launch actually newsworthy, or what angle will resonate with the media. So it defaults to the most generic version possible. It’s up to us to train it.
The Foundational Rule: Context Is Everything
The single most important principle for better AI outputs is this: give comprehensive context upfront. Think of it like briefing a new team member. The more they understand about your company, audience, and goals, the better their work will be.
Instead of "Write a press release about our new product," try:
"I need a press release for [client name], a B2B SaaS company serving mid-market retailers. We're launching an AI-powered inventory management tool that reduces stockouts by 40%. Our target audience is retail operations managers who read publications like Retail Dive and Chain Store Age. We want to position this as solving a major pain point that cost retailers $1 trillion in lost sales last year. Our brand voice is authoritative but approachable, and we avoid jargon and focus on business outcomes over technical features."
With the latter, you've told the AI who you are, what you do, what's newsworthy, who you're trying to reach, and how you communicate. The output will be exponentially better.
Structuring Your Prompts
The most effective PR prompts follow a consistent structure:
- Role and context: "You're writing for [company name], a [company description] that [value proposition]. Our target audience is [specific audience] who care about [key concerns]."
- Specific task: "Create a [deliverable type] that [specific objective]."
- Constraints and requirements: "Keep it to [length]. Use [tone]. Include [specific elements]. Avoid [things to exclude]."
- Success criteria: "This needs to [what good looks like] so that [desired outcome]."
- Framework Example: "You're pitching on behalf of [finance app], a personal finance tool for freelancers struggling with irregular income. I need an email pitch to a journalist at Fast Company who covers the future of work. The hook is new data we have showing that 68% of freelancers have missed bill payments due to cash flow gaps, despite earning above the national median income. Keep it under 150 words, conversational but data-driven. Open with the surprising statistic, explain why it matters now, and offer our CEO for an interview. The goal is to get a response, not tell our whole story in the pitch."
Power Techniques for Better Outputs
These specific techniques significantly enhance AI responses for PR, beyond just the basic structure work:
- Provide examples: If you want the AI to match your brand voice, paste in examples of previous coverage, blog posts, or pitches that captured your tone well. Tell the AI: "Match the style and tone of this example." This is especially powerful for maintaining consistency.
- Specify the audience's knowledge level: "Write this for journalists who cover enterprise software but may not understand blockchain technology" gives the AI crucial guidance on how much to explain versus assume.
- Use chain-of-thought prompting: For complex tasks, break them into steps. "First, identify the three most newsworthy angles in this product launch. Then, choose the strongest one for B2B tech media. Finally, write a pitch email using that angle." This produces more strategic, thoughtful outputs.
- Request variations: "Give me three different versions of this pitch, each emphasizing a different angle" helps you test different approaches and often surfaces ideas you wouldn't have considered.
- Iterate with feedback: Don't accept the first draft. Tell the AI what's working and what needs adjustment: "This is good, but make the opening more provocative and cut the third paragraph entirely. Also, our CEO's title is Chief Executive Officer, not President."
The key is using AI as a collaborative tool that handles the heavy lifting of drafting and ideation. At the same time, you provide the strategic thinking, relationship knowledge, and editorial judgment that make PR effective.
The Takeaway
While AI won't replace PR professionals, PR professionals who know how to use AI effectively may replace those who don't. The difference between mediocre and exceptional AI outputs isn't the technology; it's the quality of your prompts.
Treat AI like a skilled junior team member who needs clear direction, comprehensive context, and constructive feedback. Try giving it the information and structure it needs to do great work, and you'll be amazed at how much more productive your PR efforts become.
Start with one small workflow (could be a media list, industry research, or social copy adaptations) and refine your prompting technique. As you get better at communicating with AI, you'll find yourself spending less time on first drafts and repetitive tasks and more time on the strategic thinking that actually moves the needle for clients.
LEAVE A COMMENT