Most digital PR programs were built around a link. You pitch a publication, you get a dofollow link, and you count referring domains. That model still works for classic SEO, but it misses what now drives a growing share of discovery: whether AI assistants name your brand when a buyer asks a question.
AI answers are assembled from sources. When Perplexity, ChatGPT search or Gemini answers "what's the best tool for X," they typically name a handful of companies and cite a handful of pages. If your brand isn't in those pages, you're not in the answer. Digital PR is one of the few levers that changes which sources exist in the first place.
This playbook covers three steps: find out which domains AI already cites for your prompts, pitch those publications with something genuinely useful, and measure whether anything moved.
Step 1: Find the sources AI actually cites
Before pitching anyone, get the citation list. Not the list of sites you wish cited you — the list of sites that are actually showing up in AI answers for your category's questions.
Start with your real buyer questions. Write down five to ten prompts a prospect would type or say, in their words, not yours. For a project management tool that might be "best project management software for a small agency" or "how do I keep client projects on track." Avoid brand-name prompts at this stage; you want the questions asked before anyone knows you exist.
Then run those prompts through AI tools that show sources. Perplexity with web search turned on is the most transparent: it lists the pages it used. ChatGPT search shows citations too. Run each prompt more than once, since answers vary between sessions.
Our free AI visibility check is a fast starting point. It generates three buyer questions for a brand and category, asks Perplexity with web search, and shows whether the brand is named, which competitors are named, and which sources are cited. It gives you three checks per day. One caveat worth knowing before you build a process on it: it does not track over time, so it's a snapshot tool. Use it to get your first citation list, then keep your own spreadsheet if you want history.
When you have run ten or so prompts, tally the domains. You'll usually see a pattern:
- Comparison and review sites in your category
- Trade publications that cover your industry
- Reddit threads and niche forums
- Documentation, help pages and pricing pages of competitors
- A small number of listicles that keep resurfacing
That tally is your target list. It's also a reality check: if the cited domains are all review aggregators and forum threads, a pitch to a national business outlet won't move your AI visibility much.
Separate the sources you can influence
Not every cited domain is pitchable. Competitor pricing pages you can't touch. Reddit threads you can participate in, carefully — the approach in Reddit, Forums and AI Answers: Earning Mentions Without Spam is a better fit there than a PR pitch. Review platforms respond to a review-generation program, which is its own discipline and covered in How Reviews Shape AI Recommendations for Local and Product Searches.
What's left is the earned media target: journalists, editors and independent publishers who write about your category and whose pages AI keeps citing. That's where a digital PR pitch belongs.
Step 2: Pitch with data, not announcements
The publications AI cites tend to be the ones that answer questions. A listicle titled "best X for Y" gets cited because it answers a question directly. A trade piece gets cited because it explains something. A press release announcing your funding round rarely gets cited, because it doesn't answer a buyer question.
So the pitch has to offer something that makes the publication better at answering a question.
What actually gets picked up
Original data from your own business. You have operational data nobody else has: anonymized usage patterns, pricing benchmarks, support ticket themes, hiring trends in your niche. Aggregate it, remove anything identifying, and present it as a finding. "We looked at X across Y accounts and found Z" is a story. "We're excited to announce" is not.
A survey you actually ran. If you run one, report the methodology plainly: sample size, who was asked, when, how. Don't stretch the finding beyond what the data supports. A modest, honest finding is more citable than a dramatic, shaky one.
A useful free tool or dataset. Calculators, templates, benchmark datasets and public trackers get linked and cited because they're useful on their own. They also tend to keep getting cited for years.
A clear expert explanation. Sometimes a publication needs a paragraph explaining how something works, from someone who does it. Offer the explanation, not the quote. Writers can quote from a well-written explanation.
How to write the pitch
Keep it short and specific. Three or four sentences: what you have, why it's relevant to their readers right now, and what you can provide (data, charts, an interview, an exclusive). Lead with the finding, not your company.
Two practical notes:
- Match the format. If the cited page is a comparison listicle, pitch an updated comparison with your data included. If it's an explainer, pitch a fact or example that improves the explainer.
- Don't ask for a link. Ask whether the data is useful. The link follows if it is. A link you demanded and didn't earn is worth less anyway — the goal is being named in a source AI trusts, not a URL in a footer.
Where the brand name goes
AI systems name brands that appear in the source text as entities. So the mention matters as much as the link. Make it easy for a writer to name you correctly: give them the exact company name, a one-line description of what you do, and your category. If a writer describes you as "a tool for X," that phrasing is what may surface later.
Step 3: Measure before and after
PR without measurement becomes a pile of clippings. For AI visibility, you need a baseline and a repeat.
Build a simple baseline
Before you send a single pitch, record:
| What to record | Why |
|---|---|
| Your prompt list | Consistent prompts make comparisons meaningful |
| For each prompt: is your brand named? | The core outcome |
| Which competitors are named | Shows the gap you're closing |
| Which domains are cited | Your target list, and your progress measure |
| Date and tool used | AI answers change; you need to know when you looked |
Run each prompt a few times and note the variation. AI answers are not deterministic, so a single run is a weak signal. If your brand appears in two of five runs before a campaign and three of five after, that's a direction, not proof.
Track the sources, not just the mentions
Brand mentions in AI answers move slowly and noisily. Source citations are a more stable intermediate metric: either a domain you pitched now appears in the citation list for your prompts, or it doesn't. That's a cleaner read on whether your outreach is landing in the right places.
Set honest expectations
Nobody can promise a mention in an AI answer. The systems change, the retrieval varies, and a single new article may take time to be crawled and indexed. What you can control is whether the sources AI already trusts contain accurate, useful, well-structured information about your category — and whether your brand is named in it.
Give a campaign a full quarter before drawing conclusions. If you pitched five cited publications and two now cite you, that's a real result. If nothing moved, the likely cause is a mismatch between what you pitched and what those pages are for.
Common ways this goes wrong
- Pitching your news instead of their question. Editors at cited publications care about their readers' questions, not your milestone.
- Chasing high-authority domains that AI doesn't cite. A link from a famous site that never appears in your prompt citations does little for AI visibility.
- Measuring once. A single check before and after tells you almost nothing. Repeat over weeks.
- Ignoring the boring sources. Niche trade sites and small comparison pages are often cited precisely because they answer a narrow question well.
Your next step
Run your five most important buyer prompts through the AI visibility check today, write down the cited domains for each, and pick the two pitchable publications that appear most often. Then draft one pitch built around a finding only your business could produce. Send it this week — the baseline you record first is what makes the result readable later.