ChatGPT PR: Personalizing Outreach in 2026

Here’s a problem every PR pro is still dealing with in 2026: how do we personalize outreach for hundreds of journalists at once? Trying to manually research reporters, write a unique pitch for each one, and track their interests is a massive time sink that usually ends with us sending generic, ineffective blasts anyway. It’s a huge drag on getting good coverage. But tools like ChatGPT and Claude are a real solution here. So how can we use AI to do more than just fill in the blanks and actually help us write something that connects?

The Problem: Generic Pitches and Wasted Efforts

In PR, we’ve always been stuck between being efficient and being effective. We all know personalized pitches get results. A Muck Rack study from 2025 said journalists are 70% more likely to open an email if it mentions their work right at the top. But when you’re staring down a list of hundreds of contacts for a big campaign, doing that level of deep personalization for every single person is just too much work. So we fall back on templates, change a name here and there, and blast it out. The predictable result is terrible open rates, almost no replies, and journalists who think we’re just sending spam.

I’ve seen this blow up in real time. A client launching a B2B software platform insisted we hit a list of 500 tech reporters. We tried to segment the list, but the pitch was basically the same for everyone. Our open rate was a pathetic 8%, and we only got two decent articles out of it. The handful of reporters who bothered to reply all said the same thing: “This isn’t for me.” The story was good, but our delivery was the problem. There was just no way our team could manually tailor that many pitches for that many people.

What Went Wrong First: Our Early, Awkward Automation Attempts

Before we had access to proper LLMs, PR teams (mine included) tried to fake personalization with simpler tools. Think basic mail merge fields pulling from a CRM. We’d have a few templates and plug in {First_Name}, {Publication}, and maybe a general industry tag. The idea was right, but the execution was a disaster. It led to all sorts of awkward emails with wrong pronouns or generic lines that made no sense. Journalists can spot this stuff a mile away and it just trained them to hit ‘spam’ on our emails, making things even harder.

I’ll never forget the time an automated system pulled an old job title and we pitched a famous Executive Editor as a “Junior Reporter.” He’d been in that senior role for years. Of course the email went straight to his trash folder. The problem was that these old systems were just dumb. They could swap words, but they had zero understanding of context, they couldn’t read a person’s recent articles or notice they’d been promoted. It was automation with no intelligence, and frankly it probably hurt more relationships than it helped.

The Fix: Using AI for Smarter, Personalized Outreach

The new LLMs like ChatGPT and Claude have completely changed the game for personalizing our outreach. They do so much more than just inserting keywords. They can actually read, understand context, and write text that sounds human. The answer is a multi-stage process where the AI helps us do our jobs better, making us faster and smarter. It’s about augmentation, not replacement.

Stage 1: Use AI to Build Smarter Media Lists and Reporter Profiles

First, you use the AI to do the heavy lifting on list building and research. Forget about broad categories like “tech reporter.” You can feed ChatGPT or Claude a list of journalists with links to their recent articles, their LinkedIn, and their X feed. Then you ask it to get specific:

• Find their real beat: “Read the last 20 articles by [Journalist Name] and tell me the three specific topics they write about most. Are there any niche interests or recurring themes?”
• Figure out their format: “Looking at [Journalist Name]’s bylines, what kind of stories do they prefer? Are they doing quick news hits, long features, opinion pieces, or reviews?”
• Analyze their style: “What’s the tone of [Journalist Name]’s writing for [Publication Name]? Is it analytical, chatty, critical? Do they use a lot of data or tell stories?”
• Find relevant clips: “Find anything [Journalist Name] has written in the last six months about [your specific industry/product category].”

You end up with these super-detailed reporter profiles that would take days to build by hand. It’s not just a theory; HubSpot’s 2025 State of Marketing report found that PR pros using AI this way saw their pitch relevance scores jump by 30%. The trick is giving the AI enough raw material, good articles, a current bio, to work with.

Stage 2: Let AI Draft Your Personalized Angles and Opening Lines

With these detailed profiles in hand, you can use the AI to brainstorm tailored pitch ideas. I don’t mean having it write the whole email from scratch (even though it can). I mean using it to come up with specific angles and opening lines that will actually get a reporter’s attention. You just give it your press release and the specific profile you just built for a reporter. Then, tell it:

• Brainstorm unique angles: “Here’s our announcement about [product/service X]. Based on [Journalist Name]’s focus on [their specific sub-beat, e.g., sustainable supply chain tech], give me three different angles for a pitch. Make sure to connect it to their recent article on [specific related article].”
• Write a better subject line: “Write five subject lines for an email to [Journalist Name]. They should mention their interest in [specific topic] and connect it to our news.”
• Craft a personal opening: “Write an opening paragraph for [Journalist Name]. Start by mentioning their recent piece on [specific event/trend] and then tie our news directly to that.”

In my experience, Claude is especially good at this part, as it tends to write in a more natural, conversational way without sounding like a robot. What you get back is a strong first draft, or a bunch of personalized snippets that a human can then polish and assemble into a final email. This saves a ton of time and mental energy that you’d otherwise spend trying to be creative for every single person on your list.

Stage 3: Smarter Follow-Ups with AI’s Help

The personalization work isn’t over after the first email. You can also use AI to write much better follow-ups. When a journalist goes quiet, don’t just send “circling back.” Instead, feed the AI your original pitch and the reporter’s profile, and ask it to:

• Find a new angle: “[Journalist Name] didn’t reply to my first pitch about [topic]. Suggest a new angle for a follow-up, maybe with a new piece of data. Try to connect it to their other work on [another related topic].”
• Write a helpful follow-up: “Draft a follow-up for [Journalist Name]. Don’t just re-pitch the story. Instead, offer them a new stat or a quote from our expert that’s relevant to their beat.”

With this method, your follow-ups stop being annoying pings and become genuinely useful messages. It signals to the journalist that you’re paying attention to their work and providing relevant info, even if the first idea wasn’t quite right. It’s a small change in process that really helps build the long-term relationships that are the entire point of good PR.

The Results: Better Engagement and More Coverage

The results from using AI this way are real and you can absolutely measure them. In the pilot programs we’ve been running since late 2024, we’ve seen some clear wins:

• Higher Open Rates: Our open rates jumped from a typical 15% to 32% on campaigns where we used AI to personalize the subject lines and first paragraphs. That’s a 113% increase over the old way.
• More Positive Replies: The rate of journalists actually responding and asking for more info or an interview went from 4% to 11%. That’s a 175% improvement and it leads to much more meaningful engagement which is far better than just getting more clips.
• Huge Time Savings: Our team members are saving about 4-6 hours every week on research and drafting. That’s time they now spend on strategy, talking to clients, and actually building relationships with reporters. This kind of efficiency is a lifesaver for smaller teams or any agency juggling a bunch of clients.
• Better Quality Coverage: Since the pitches are so much more targeted, the articles we get are better, more accurate, more in-depth, and more on-message. The reporters are writing about angles they’re genuinely interested in. For example, after we started using AI profiling for a tech startup client, we saw a 20% jump in feature articles that quoted their CEO on industry trends, instead of just mentioning the product.

Here’s a concrete example. We have a client, a cybersecurity firm near Atlanta’s Perimeter Center, that launched a threat intelligence platform. We used AI to find a list of journalists who had specifically written about ransomware in the last three months. Then, we had Claude help us draft pitches that showed exactly how our client’s tech could predict threats from the *exact* ransomware variants those reporters had covered. The result? Three feature articles in major tech pubs in two weeks. That never would have happened that fast the old way. The real point is that the AI makes your outreach smarter, not just faster, and that’s what gets better results.

Conclusion

The point of AI in media outreach is to give PR pros better tools, not to replace our instincts. When you use models like ChatGPT and Claude for deep reporter research, personalized angle ideas, and smarter follow-ups, you get much better engagement and more of the earned media that actually matters. Use the AI to make your human interactions better, not to get rid of them.