The Media Copilot

The click from AI search isn't the finish line. It's barely the starting gun. Publishers focused on converting AI referral traffic are solving half the problem. The more important work happens earlier, inside the AI conversation itself, where readers are already forming opinions about which sources to trust. [Pete Pachal](https://lnkd.in/eAN9DXaZ) breaks down what that journey actually looks like, drawing on new research that segments AI users by intent: orientation readers getting up to speed, evaluators comparing perspectives, action readers ready to commit, and support readers who already have. Each group wants something different, and each requires a different content response. The strategic implication is significant. Even when a citation doesn't produce a click, it can still shape how a reader perceives your publication as they move through their research. Visibility across the journey matters, not just at the moment of conversion. The publishers best positioned to grow in AI search won't just optimize for clicks. They'll build content that earns citations at every stage of reader intent. https://lnkd.in/ebu6QXtF #AISearch #DigitalPublishing #MediaStrategy #ContentStrategy #AudienceDevelopment

Most conversations about AI in local news stay at the level of promise. Axios is already in the operational details. The company now runs local newsrooms in 35 cities, with eight more coming by year's end. Some of those markets have one reporter. Some have half a reporter. The model only works if AI is doing real work behind the scenes — social publishing, newsroom training, editorial planning — while human journalists stay focused on original reporting. On the latest Media Copilot podcast, Axios COO Allison Murphy is candid about why this matters: the fundamental challenge facing local journalism is financial, and the window to solve it is narrowing. Tools like the "Axiomizer" and "Localizer" aren't side experiments. They're part of a deliberate strategy to bring down the cost of quality journalism without gutting what makes it credible. The harder question Murphy addresses is where to draw the line. Audience trust, AI transparency, and editorial standards all come up — and none of them get hand-waved away. For anyone thinking seriously about the future of local news, media sustainability, or what AI adoption actually looks like inside a working newsroom, this is worth your time. Listen to the conversation here:

Journalism's relationship with AI was finally showing signs of progress. Then a plagiarism scandal at The New York Times reminded everyone how quickly that progress can unravel. A freelance book review submitted to the Times contained passages nearly identical to an earlier Guardian review of the same book. The writer admitted using AI. The Times cut ties. And suddenly, every newsroom that had been cautiously warming to AI tools had a fresh reason to pump the brakes. The incident is worth examining closely, because the failure wasn't really about AI. It was about how the AI was used. Evidence suggests the tool pulled from web search to generate copy, likely surfacing the Guardian review without the writer explicitly asking for it, and without any prompt guardrails telling it not to. That's a process failure, not a technology failure. [Pete Pachal](https://lnkd.in/eAN9DXaZ) argues the path forward isn't to retreat from AI adoption, but to get serious about how it's deployed: prompt design, private testing, clear policies, and genuine accountability for the outputs you publish under your name. The stigma around AI in journalism hasn't disappeared. Incidents like this feed it. But so does vague, undisciplined use of powerful tools. https://lnkd.in/eQYwpHhq #AIinJournalism #MediaIndustry #ArtificialIntelligence #NewsroomInnovation #ContentStrategy

The platforms had the tools to flag AI-generated fakes. They just didn't use them. A 22-year-old medical student in India built a fictional MAGA influencer named Emily Hart using Google Gemini, grew her to millions of views across multiple platforms, and pulled in thousands of dollars a month from an OnlyFans competitor and merchandise. Not a single AI content label flagged what audiences were actually looking at. This isn't a story about one clever fraud. It's a story about a system that was supposed to protect users from exactly this kind of deception — and quietly failed. Content Credentials and similar standards have existed for years. Platforms implemented them. Then made them nearly impossible to find. Meanwhile, the rest of this week's issue covers a lot of ground on how AI is reshaping trust in media: 🔹 Major publishers and Scott Turow are suing Meta, alleging its Llama models were trained on millions of copyrighted books without permission 🔹 A UN Women report found 45% of women journalists now self-censor on social media due to AI-enabled harassment — up 50% since 2020 🔹 An AI podcast about the Epstein files hit 2 million downloads with no human editor involved — raising real questions about accountability Pete Pachal breaks down why the labeling infrastructure exists, why it isn't working, and what that means for anyone trying to tell real from fake online.

Most comms leaders have spent years explaining why their work matters. AI may have just handed them the proof. Gartner's 2026 Communications Predictions include a striking data point: more than 95% of citations inside AI-generated answers are non-paid. Earned media accounts for roughly 27% of those citations. Paid media barely registers. That means the channel most people will soon use to learn about a brand is almost entirely shaped by journalism, third-party coverage, and expert commentary — the work comms teams have always done but rarely been able to quantify. The discipline emerging around this is called generative engine optimization (GEO), and it reframes what PR is actually for. The metrics it produces — citation share, sentiment inside AI answers, narrative alignment across generative surfaces — are ones a CFO can actually read. Comms stops looking like a cost center and starts looking like the team that owns a measurable, strategic asset. The catch is a real one. Most communications functions aren't built for this yet. The skills required span earned media strategy, content architecture, data analysis, and AI literacy. Teams that develop this fluency will own something genuinely valuable. Those that don't will find traditional metrics look even weaker against the new benchmarks. Pete Pachal breaks down the full strategic picture, including what the budget data reveals and why this moment is different from previous "PR is back" cycles:

The idea sounds simple: AI remixes any story into any format, on demand. The reality is messier. "Liquid content" is the concept making the rounds at industry conferences right now, and the demos are genuinely impressive. Google NotebookLM turns a folder of documents into a podcast. Amagi scans a live newscast and generates short-form videos in real time. Stringr pulls licensed footage to turn a news article into video, automatically. What the demos tend to skip: the complications that show up the moment you try to run this at scale inside an actual newsroom. A few things worth thinking through: 🔹 Repurposing content isn't new, but AI changes the cost and speed equation significantly 🔹 The tools are real and showing up on conference floors, not just in pitch decks 🔹 The gap between "it works in a demo" and "it works in a workflow" is where most implementations stall 🔹 Treating liquid content as a growth engine without understanding its limits is where publishers get into trouble Pete Pachal breaks down why the theory is compelling, where the friction actually lives, and what media companies need to understand before building around this idea.

Thrilled to be featured on News Media Help Desk this week, where I share some of my secrets for journalists interested in using deep research to level-up their work. One of the best use cases is using the tools to quickly find subject matter experts. By spending just a few minutes refining a good prompt, a deep research tool can spin you up a list, complete with biographical data and contact information, in minutes. This works in comms, too: creating media lists can often be a big time suck in PR, but deep research can help find not just the right list of reporters, but influencers and creators, too. What's your top use case for deep research? I'd love to know how folks are using it, especially now that you can target it at things like Google Drive folders, your inbox, and more.

B2B media isn’t competing for clicks anymore. It’s competing for decisions. In the latest episode of The Media Copilot podcast, Pete Pachal sat down with Keith Turco, CEO of Madison Logic, to unpack how AI is reshaping B2B marketing—and what that means for publishers. The shift is pretty stark: 🔹 Buying groups now include 5–16 stakeholders 🔹 Each one needs multiple touchpoints before a decision 🔹 And AI is compressing the distance between question and answer That combination changes the math. Scale matters less. Precision matters more. Turco’s core argument: B2B media needs to move from “one-to-many” to “one-to-few.” That applies to everything: 📣 Messaging 🎯 Targeting 📊 Measurement And especially how value is defined. Because in B2B, ROI isn’t one thing. For some, it’s pipeline. For others, it’s appointments. For others, it’s shortening a sales cycle. Publishers that can’t map their offerings to those outcomes are already behind. There’s also a bigger mindset shift happening. AI isn’t replacing marketing—it’s refining it. It helps: 🔹 Identify buying groups 🔹 Personalize messaging at scale 🔹 Connect fragmented media touchpoints But it doesn’t remove the need for strategy. If anything, it raises the bar. One more thing: Turco defines the “at-work state of mind,” meaning decision-makers aren’t just at their desks anymore. They’re: listening to podcasts on a treadmill,s crolling social feeds on weekends, and moving fluidly between personal and professional contexts That’s why formats like audio—and yes, podcasts—are becoming central to the B2B mix. The takeaways: 👉 Audience quality > audience size 👉 Measurability > impressions 👉 Relevance > reach And increasingly: 👉 Performance > everything B2B media still matters. But only if it plugs directly into outcomes marketers can prove. 🎧 Listen or watch on... YouTube: https://lnkd.in/eKHHzsTx Spotify: https://lnkd.in/enBspb78 Apple: https://lnkd.in/ewSstec7

Most publishers are treating AI search traffic as one big bucket. It isn't. The people who arrive at your content through ChatGPT or Perplexity aren't a uniform audience. They're in four distinct stages of intent: orienting themselves on a topic, comparing options, looking to take a specific action, or ready to decide. Each group behaves differently inside the chatbot, and each needs a different content strategy to convert. The conventional wisdom says AI referrals are higher quality than traditional search clicks. That's true as far as it goes. But quality traffic that lands on the wrong content for its intent stage still bounces. Understanding the journey before the click is what separates publishers who build AI search loyalty from those who just get occasional visits. This week's issue maps out that intent framework and explains what publishers can actually do with it. Also inside: 🔹 GPT-5.5 launches with benchmark gains and roughly half the compute cost of competing models 🔹 AAM opens its Ethical AI Certification to all member publishers as audience trust concerns sharpen 🔹 Gist GEO raises $75M to track brand visibility inside AI-generated answers 🔹 Tubi puts streaming directly inside ChatGPT 🔹 ChatGPT Images 2.0 takes a serious run at rendering legible text in generated graphics Pete Pachal breaks down the full AI search funnel and what publishers should do differently at each stage:

Our GEO Dinner is Tuesday, April 28 in New York. A handful of seats are left. It's an intimate evening for senior media and communications leaders working out what generative engine optimization actually means for their organizations. A focused briefing from Amanda Coffee and Pete Pachal, dinner with peers facing the same questions, and takeaways you can bring back to your team. No panels. No sales pitches. Seats are limited, sp grab one of the last tickets while you can👇 https://lnkd.in/eBpKwNb4 #GEO #AEO #AIcomms #AImedia