From building a journalist list to reading the sentiment after the story runs: here's the AI-assisted media outreach workflow PR people are actually using in 2026.
I use AI to speed up media list building, personalize pitches at scale, catch journalist requests on platforms like Qwoted and Featured/HARO, and monitor coverage and sentiment after we land a placement. The relationship-building and judgment calls stay mine. Journalists can tell the difference, and 88% of the 1,000+ reporters Muck Rack surveyed for its 2026 State of Journalism report say they’ll reject a pitch that doesn’t match their beat, AI-written or not.
Every PR campaign I’ve run starts in the same place: a list of the right people to contact. Get that part wrong and the quality of your pitch stops mattering much. You end up emailing a tech reporter who moved to a lifestyle beat eight months ago, or a freelancer who hasn’t published in the outlet you think they still work for. I’ve done both of those, early in my career, and the bounce-back email is its own special kind of embarrassing.
For years, building that list meant hours inside a media database, exporting spreadsheets, and guessing which contacts were still accurate. AI has actually changed this part of the job. When I first ran Muck Rack’s list agent on a client search last spring, I expected a slightly faster spreadsheet. What I got was a list with fewer dead ends on it, which turned out to be the more useful thing.
Muck Rack launched a Media List Agent in beta in February 2026 that builds media lists from a plain-language prompt describing who you’re trying to reach. It surfaces up to 50 recommended journalists with an explanation for each match, tied to their recent coverage, beat, and location, and it will also review a list you built yourself and flag who to add or cut based on your actual pitch content. In beta testing, Muck Rack reported customers using the agent saw an average 10% increase in pitch open and click rates, largely because fewer of those pitches were landing on the wrong desk. Most of that lift, from what I’ve seen on our own lists, comes from cutting the people who were never going to open the email in the first place.
Prowly, now operating as the Semrush AI PR Toolkit after Semrush’s acquisition, gives you an AI-Cited Media Database with over a million journalist and outlet contacts, filterable by beat, geography, and outlet type, plus a newer angle: it flags which outlets are actually getting cited by AI answer engines like ChatGPT and Perplexity in your industry, so you can prioritize coverage that shows up beyond the article itself. I’ll admit that feature didn’t seem like much when I first read the launch notes, but it’s become one of the first things I check now when a client asks where their coverage should go.
Honestly, the biggest mistake I see non-PR marketers make here is trusting an AI-generated list without opening a single article. The tool did the research. You still have to read three sentences of it before you hit send.
Here’s the number I tell every client before they write a single pitch: the average response rate across cold media pitches is 3.43%, according to Propel’s State of PR research. Only about 8% of pitches sent ever turn into published coverage. AI doesn’t touch that math on its own. What it can do, and this is the part I actually care about, is help you stop being part of the 86% of pitches journalists say they reject outright for not matching their beat.
Use AI to draft a first version of a pitch fast, then spend your actual time on the part that moves the needle: personalization. I feed the model 2 to 3 recent articles from the specific journalist and ask it to draft an angle that connects the news to a pattern they’ve already been covering. It’s not perfect on the first pass, the tone usually reads a little stiff, but it gets me a starting point in five minutes instead of forty. Muck Rack’s own 2026 research found personalized, beat-matched pitches get a meaningfully higher response than templated ones, and that a short pitch, ideally under 200 words, outperforms a long one by a wide margin.
This is a natural companion step to writing the release itself. If you haven’t put together the actual document yet, our guide on how to write a press release with AI covers that piece in detail. This article picks up from where that one leaves off, everything that happens around the release: list, pitch, and placement.
Journalists can smell a mass AI pitch from the subject line, honestly. Muck Rack’s 2026 report found 38% say AI-generated pitches are easy to spot and usually get deleted on sight. I had a reporter reply to a pitch once, not out of interest, just to tell me it read like it had gone out to six other outlets that week. She wasn’t wrong. The tell isn’t grammar, it’s genericness. If your pitch would work for any outlet in your category, it’s not personalized enough, no matter which tool wrote the first draft.
Beyond outbound pitching, there’s a whole reactive side of PR: journalists actively looking for sources right now, on a deadline, for a story that’s already assigned. This is the HARO-style workflow, and it changed enough over the past two years that I had to relearn part of my own process.
HARO (Help a Reporter Out) was discontinued by Cision in late 2024 after being rebranded as Connectively, then sold to Featured.com in 2025, which revived it as a free, ad-supported, email-based service. It’s back to its original model: journalists post requests, you reply if it fits, up to three digests a day, no cost. Featured says the network now connects more than 800,000 sources with over 75,000 journalists.
Featured.com layers AI on top of that free stream through what it calls PR Intelligence, a feature on its paid Business Plan ($59 to $99 per month per profile) that shows you how many answers a request has already received and how many the journalist has selected, so you’re not pitching blind into an opportunity that’s already been filled. I like this feature more than I expected to. It’s saved me from drafting a full response to a request that already had thirty answers sitting in front of it.
Qwoted runs a parallel marketplace model: journalists post requests privately (they’re hidden until an expert responds, which cuts down on spam), and Qwoted uses matching to surface relevant requests to the right experts. It’s free with limits, or $99 a month for the unrestricted Professional plan.
Industry research on journalist-request platforms suggests the overlap between any two of them runs around 17%. In plain terms, each platform sees a mostly different set of requests. If you’re only watching one inbox, you’re missing most of what’s out there. When I set up a media-request routine for a B2B client last year, we ran HARO and Qwoted side by side for a month, and in practice the overlap felt even thinner than that. This is one place where an AI layer earns its keep: instead of manually scanning three or four digests a day, a tool that watches multiple sources and flags only the requests that match your actual expertise saves real hours.
Landing the placement is only half the job. The other half is knowing it happened, where, and whether the coverage actually reads the way you wanted it to. I learned this the hard way early in my career, tracking placements in a spreadsheet that was already out of date by the time I opened it on a Monday morning. This is where AI-powered media monitoring earns its budget line.
Brand24 is built around AI sentiment analysis and topic detection across news, social, and forum mentions, and it’s priced accessibly for small teams starting around $29 a month. Meltwater operates at enterprise scale, ingesting more than 1.3 billion documents a day across 240-plus languages, with sentiment scoring layered across media monitoring, social listening, and influencer tracking in one platform, priced on a custom quote. Determ sits in between, with real-time tracking plus an AI assistant called Synthia that generates plain-language summaries of your most influential coverage so you’re not reading fifty articles to find the three that matter.
If your coverage strategy already touches organic social conversation, it’s worth pairing this with the broader listening stack we cover in our guide to AI social media analytics tools, since a lot of media monitoring platforms now blend earned coverage and social mentions into a single sentiment view.
A word of caution here, from someone who’s been burned by trusting a dashboard too much: AI sentiment scoring is directionally useful, not gospel. It reliably tells you when coverage skews negative versus neutral versus positive across dozens of mentions faster than a human could read them all. It’s much less reliable at catching sarcasm, nuance, or a technically positive headline sitting over a critical article. I once had a tool flag a piece as neutral that was, read in full, pretty clearly needling the client. Skim the actual top pieces yourself. Don’t just trust the score.
PR has always struggled with a specific problem: proving that a pitch caused a placement, and that the placement was worth the effort. Spreadsheets and manual tracking used to be the only option, and most teams let this slide until a leadership review forced the issue. I’ve sat in enough of those reviews scrambling to rebuild a timeline of who sent what to know how much time this used to waste.
Muck Rack’s newer Dashboard Widgets attach Pitch Placement markers directly to reporting, so a pitch that leads to coverage shows up connected to that outcome automatically, without someone manually matching a sent email to a published article days or weeks later. It’s a small feature, but it solves a genuinely tedious task, one I used to do by hand on a Friday afternoon.
There’s also a newer wrinkle worth tracking if you’re reporting PR results to anyone above you: some of these platforms, Muck Rack included, are starting to show whether a piece of coverage also shows up as a citation in AI-generated answers from tools like ChatGPT or Perplexity. A quote in a trade publication used to only reach that outlet’s readers. Now it can also become part of how an AI model answers a question about your industry months later, and that’s changed how I think about pitching a smaller, more topically focused outlet over a bigger generalist one.
Here’s the part most vendor pages skip. According to Muck Rack’s 2026 State of AI in PR research, only about 12% of PR professionals say they’re actually using AI agents today, well below the hype you’d expect from how much these platforms talk about AI in their marketing. I don’t think that gap is just slow adoption. I think it’s PR people correctly sensing where the tools stop being useful, because I’ve felt those same edges in my own work.
AI is good at pattern matching across large volumes of contacts, articles, and mentions. It’s bad at the parts of PR that were never really about volume. Reading a journalist’s tone in a two-line reply and knowing whether that’s genuine interest or a polite brush-off is one of those parts, and I haven’t found a tool that gets it right consistently. Same with deciding whether a story angle is actually newsworthy or just convenient for your client. And then there’s the slower work of building the kind of relationship where a reporter calls you first when they’re working on something in your space, which takes longer than any software subscription.
Only 30% of journalists in Muck Rack’s 2026 survey said PR relationships are very important to their success, but 86% said at least some of their stories start with a PR pitch. I read those two numbers as good news, honestly. Relationships still matter, they’ve just stopped being the only lever you can pull. AI handles a lot of the volume and research side now, which frees up time for the relationship side. It doesn’t remove the need for it.
If you’re evaluating whether a piece of AI-drafted pitch copy or press material is actually ready to send, run it through the same scrutiny you’d apply to any AI-generated marketing content before it goes out under your name. Our guide to evaluating AI-generated marketing content walks through exactly that check before something goes live.
You don’t need every tool in this article. What you need depends on whether you’re doing PR solo, running it inside a small marketing team, or managing an agency book of clients, and I’ve built a version of this stack at each of those scales over the years.
Start free. Sign up for HARO through Featured.com to catch journalist requests at no cost, and add Qwoted’s free tier as a second source, since the two platforms don’t overlap much. Use whatever AI assistant you already have (ChatGPT, Claude, or Gemini) to draft and personalize pitches by hand. Skip paid media databases until you’re pitching regularly enough to justify the spend. This is roughly where I started, back when I was running campaigns one client at a time.
This is where a paid media database starts to pay for itself. Prowly’s AI PR Toolkit tier or Muck Rack’s core plan both handle list building, pitching, and basic monitoring in one place. Add Featured’s Business Plan ($59 to $99 a month) if journalist requests are a real source of coverage for you, since PR Intelligence saves you from wasting pitches on nearly-filled requests.
At this scale, monitoring and reporting matter as much as pitching. Pair a full media relations platform (Muck Rack or the Semrush AI PR Toolkit) with a dedicated monitoring tool like Meltwater or Determ so client sentiment reporting doesn’t eat a full day every month. Budget for enterprise pricing here; these tools are quoted individually and scale with the volume of mentions and seats you need.
One honest note on budget, and I say this after wasting money on at least one tool that overpromised: every option in this article has a free tier, a trial, or a demo. Test the actual list-building or monitoring output on your real beat before you commit to an annual contract. The demo data always looks better than what shows up once the tool is working against your specific industry.
Yes, and honestly I still run the free version for most of my own client requests. Cision discontinued the original HARO service in late 2024 after rebranding it as Connectively, then sold it to Featured.com in 2025. Featured revived HARO as a free, email-based service connecting more than 800,000 sources with over 75,000 journalists, and layers an AI-assisted co-pilot on top of it through its paid plans for people who want to do more than scan free digests.
It depends on your budget, honestly. Muck Rack’s Media List Agent and Prowly’s AI-Cited Media Database (now part of Semrush’s AI PR Toolkit) are the two most established options, both letting you build a list from a plain-language description of who you’re trying to reach. Neither replaces manually checking that the journalists are still active on the beat you’re pitching, a step I never skip.
It can. Muck Rack’s 2026 State of Journalism report found 38% of journalists say AI-generated pitches are easy to identify and usually get deleted. The giveaway is usually genericness, not grammar, in my experience. Draft with AI if it saves you time, but personalize the angle by hand using the journalist’s actual recent work before you send it.
A press release is one document in a much bigger workflow. This guide covers everything around it: building and qualifying who receives it, personalizing the pitch that goes with it, catching reactive journalist requests, and tracking what happens to your coverage after it runs. If you need the release itself, our dedicated guide on writing a press release with AI covers that step specifically.
You can, and I’d argue most solo marketers should start there. Run a lean version of this entire workflow with a general AI assistant plus free tiers of HARO, Qwoted, and a basic monitoring tool like Brand24’s starter plan. Paid PR platforms earn their cost once you’re managing dozens of contacts and multiple active pitches at once; below that volume, manual tracking with AI-assisted drafting works fine.
I checked this against Muck Rack’s own 2026 State of Journalism report and its February 2026 Media List Agent launch data, Propel’s State of PR pitch research (via AMW’s 2026 benchmarks), and the current product pages for Prowly, Qwoted, Featured.com, Brand24, Meltwater, and Determ, rather than pulling from secondhand roundups. Where a number came from a vendor’s own launch data, like the 10% lift from Muck Rack’s list agent, I’ve said so directly rather than presenting it as independent. The workflow itself is the one I actually run for clients: build the list, draft and personalize the pitch, watch the journalist-request platforms, and track what happens after the story runs. Every stat and tool claim below is sourced and linked.
Weekly tips. Real examples. Practical help for busy professionals.
We care about your data, check out our privacy policy.