Amanda Coffee is CEO of Coffee Communications and ex-Under Armour, PayPal and eBay.
AI is no longer a fringe beat. It is a core operating concern for newsrooms and the communications teams that work with them. That shift is changing what makes an AI story worth covering and how it needs to be pitched.
This Q&A features Pete Pachal, founder of The Media Copilot, a media platform reporting on how AI is changing media, journalism, and content creation. Pachal helps newsrooms and communications teams adopt generative AI with rigor, ethics, and measurable outcomes. He writes, teaches and consults on what is signal and what is noise in AI and media.
Previously, Pachal led editorial and strategy at Mashable and CoinDesk, where he built teams, launched products and grew audiences at scale. His perspective reflects how newsrooms are actually using AI today. And what PR leaders need to understand to be relevant in 2026.
Q: What advice do you have for PR professionals who have a client at the intersection of AI and media, specifically what to consider before pitching you?
A: I’d separate the things I’m most interested in into three buckets:
- Ideas, products, and platforms that address the massive strategic shift that’s underway in our media ecosystem. Fundamentally, people are increasingly getting their information from AI search engines, chatbots, and other portals. That “disintermediation” effect is forcing a broader rethink in how media companies operate. It’s both a huge strategic challenge and an opportunity. I’m interested in novel approaches that media operators, journalists, and others can take in response.
- New tools that journalists, PR pros and content creators can leverage to either speed up or augment the work they do.
- Ideas that address the pain points around ethics, privacy and legal issues that arise as the transition to AI continues, especially copyright.
Finally, as an AI enthusiast and trainer, I don’t frown on PR pros using AI in their process, but when pitching, it’s important to have a human touch, even if it’s “humanizing” the work of AI. In other words, AI is OK, but slop isn’t.
What AI experiments in newsrooms look promising, and which look risky?
I’d say AI tools that accelerate investigations are the most successful implementation of the technology from an impact-to-tradeoffs standpoint. They can allow journalists to analyze massive document troves — far larger than humans alone could hope to process — and because humans are filtering everything, there’s no risk of AI tainting the final copy.
The riskiest projects are those that use AI-generated content for public-facing copy or media that’s applied at scale, meaning humans aren’t in the loop to vet every generated response. They could be chatbots like The Washington Post’s Ask the Post AI or Politico Pro’s AI-generated research reports. That’s always going to carry some risk, because generative AI hallucinates some nonzero amount of the time. That said, you can mitigate a good chunk of the risk with good product design and practicing “radical transparency” about the AI’s limits.
Where is AI helping journalism quality instead of hurting it?
Besides the investigative tools I mentioned, AI is augmenting the skill set of the “everyday” journalist. With agents, a reporter no longer needs to be, say, a legal expert to find and analyze court documents, at least on a first pass. AI-powered data analysis means any reporter can be a data journalist. And vibe coding enables journalists to create visual and interactive stories—or parts of stories — without needing to involve their product teams. If AI helps you explain the news better, get stories out faster and distribute them in more places, that’s a win.
What were some of your most popular stories of 2025? Why did those stories spark debate and social sharing inside the media industry?
My most viral post by far was actually a piece I did on the news about The New York Times embracing AI tools for its reporting staff. Like it or not, the Times still sets the standard in American journalism, and people are intrigued not just in what it reports, but how. Many were interested in what it was doing, but quite a few were dismayed, as you might imagine. And there is clearly a deep hunger for guidance on GEO, or generative engine optimization — the idea of optimizing content so it gets summarized by large language models. My most popular column for Fast Company was focused on that topic.
The Media Copilot has expanded to include a podcast, how-to section and AI trainings, platforms that build community for media professionals interested in AI. What are some recent highlights you’ve seen from your community?
Every time I close one of the longer six-week programs I teach, the final class is a capstone project review showcase, where the students show the projects they’ve built over the duration of the course. They continue to blow me away with their ingenuity. In our most recent AI for PR & Comms cohort, students build really impressive tools for media monitoring, replacing much of the grunt work with essentially agentic AIs. And in our AI for Journalists cohort, a student built an insanely detailed dashboard to help with image authentication — using content credentials to quickly verify an image’s history.
What PR pitch did you receive recently that was an immediate ‘yes’ and why?
I was intrigued by the Time AI Agent when it debuted, since it represented not just another chatbot, but a potential new business model, where media companies might “agentify” their archives to retain more control over the experience. So when their PR reached out and suggested interviewing their COO about it as well as the publication’s broader AI strategy, I jumped at the chance.
Any PR pitches that fell flat. What was missing?
Too many! Besides the pitches that were merely off-target, I think getting timing right in AI is extremely difficult, because the industry moves so fast. For example, a large free-image service pitched me on its own flavor of generative imagery, and the idea was pretty solid, but once Gemini’s Nano Banana Pro came out, everything that came before was more or less obsolete.
What do you wish more PR professionals knew about how newsrooms operate in 2026?
Newsrooms and even sections today need to be much more focused. If a pub was in a niche, it’s now even more niche. If they’re local, they’re even more local. Everyone is trying to cultivate deeper, direct relationships with their audience, and what you pitch — even how you pitch — needs to align with that goal. Keep asking yourself: How can I help this or that reporter tell a better, richer story that will keep his audience coming back? Depth and impact are in, reach and “quick hits” are out.
AI’s impacts on media changes week to week. Looking into 2026, what topics are you looking to cover more?
GEO, definitely. I hope for some movement on the broader copyright issues. And I’d like to see more clarity on the potential of the bot internet that’s emerging beside the “human” one. Can that activity be monetized in a way that might grow? I’m excited to find out!
