The Weekly Bulletin | October 28, 2025

Catch up on your members' content, check out the community buzz, and browse through job opportunities

Author

Vahe Arabian
October 28, 2025

Hi SODP community,

Let's recap on what's been happening, the new content, industry updates, tips, and more.

.TIP OF THE WEEK.

Publishers are rolling out agents – now comes the hard part

Agents aren't a demo anymore; publishers have them running real jobs. Today it's RFP interpretation, planning, pacing, and curation. Next up: direct pipes, fewer hops, and a smaller take-rate.

The current programmatic stack evolved into exactly what it promised to avoid: opacity, inefficiency, and value extraction. Publishers watched as 50 percent cuts became standard, with ads hopping through SSPs, DSPs, verification layers, and data brokers before reaching audiences. Each hop added latency, reduced transparency, and stripped revenue. The complexity made attribution nearly impossible and debugging a nightmare. Now, with agentic AI, publishers have a chance at a do-over.

The promise is enticing: agentic AI-driven media trading could wipe out many of the problems caused by its programmatic predecessor. Namely, ad tech middlemen taking cuts as high as 50 percent. Instead of having ads pass through countless vendors, agents would effectively cut the line, so more money reaches the publisher, and results are easier to see.

The end goal: a seller and a buyer agent coordinating directly, with ideally no coordination agent in the middle. What publishers don't want: the recreation of a multi-step programmatic ad tech stack.

As one revenue exec at a major publisher put it: "Programmatic is really due for an upgrade." Agentic-driven media buying could be it.

Around 50 percent of Permutive's 150+ publisher clients are using agents in day-to-day operations like RFP interpretations and campaign planning. The company's agents are already live across the direct buying cycle from audience management to optimization and reporting.

Here come the caveats

As one publishing exec put it, the first seller agents are like interns: eager, useful, but needing supervision. No one hands sales to an intern for a reason: they underprice, grab the quick win, and skip guardrails. Early buying agents will behave the same way.

"Everything that any junior revenue sales person in the entire media ecosystem has ever done wrong is going to happen at scale with agents. You need to walk into this with your eyes open, knowing it's going to fail," they said.

Another reason to cut intermediaries: fewer hops mean faster fixes when things break. "If we have multiple hops, it's all going to go horribly wrong, because trying to coordinate across multiple companies is hard, as well as multiple companies using black box agents – it's going to be a mess," said the same exec.

Regardless of adoption protocols like AdCP, agent adoption will only happen at scale if outcomes drive solid results for buyers and drive up yields for publishers. As one industry exec stressed: "Transactional volume is not there today, and it won't move there unless our [publisher ad] products outperform Facebook and Google. That's where publishers need to focus – on building high-quality ad products that deliver outcomes."

Practical steps for publishers:

  • Start small, one-to-one: Pilot direct buyer-seller agent partnerships with willing agencies. Avoid recreating the multi-hop programmatic stack.

  • Set guardrails early: Treat early agents like supervised interns – establish pricing floors, approval workflows, and quality controls before scaling.

  • Focus on product quality: Agent infrastructure won't matter if your ad products don't outperform walled gardens. Prioritize outcomes over adoption speed.

  • Prepare for failure at scale: Early mistakes are inevitable. Build monitoring systems and fast-fix protocols for when agents underprice or miss guardrails.

.SODP EVENTS.

PubTech 2025: Smarter Workflows, Safer Platforms, Stronger Connections

November 17-19, 2025 | Virtual Event | Free Registration

Join 300+ digital publishing professionals for the 3rd annual PubTech virtual event. As AI moves from experimentation to embedded infrastructure, and search engines roll out generative answers that threaten traditional discovery, publishers need practical strategies for 2026.

Three focused days:

  • Day 1 (Nov 17): Next-Gen PubTech trends for 2026 + Automation workshop

  • Day 2 (Nov 18): Brand safety in the synthetic content era + Privacy as competitive advantage

  • Day 3 (Nov 19): Zero-click world optimization + Designing future publishing stacks

Featured speakers include Anabelle Nicoud (IBM AI News & Editorial Strategy) and Eric Ulken (Stanford University, former VP Product at The Baltimore Banner). Past attendees include teams from The Guardian, Financial Times, The Economist, Indian Express, and New York Post.

4 PM & 5 PM CET | 10 AM & 11 AM ET - Register for single sessions or the full event.

PubTech Dinner London: The Product & Technology Innovation Series

November 17, 2025 | 6:30 PM | London

An exclusive off-the-record dinner for senior product, engineering, and technology leaders on the opening night of PubTech 2025. Following successful SODP-hosted dinners in New York and Dubai, this intimate gathering explores how smarter workflows, safer platforms, and stronger connections are shaping publishing's future.

Guest speaker Mel McVeigh (former Condé Nast, global media advisor) will share insights on uniting product and design disciplines to create scalable transformation roadmaps for publishers.

What to expect:

  • Thought-provoking discussions on workflow automation, platform security, and cross-functional collaboration

  • Peer exchange with senior product and tech leaders

  • Practical frameworks for integrating AI into product design and development

  • Continued networking through private WhatsApp group

Location: Cornus Restaurant, 27c Eccleston Pl, London SW1W 9NF

Seats are limited. Partner: Multidots

California Dinner: The Audience & Revenue Innovation Series

November 13, 2025 | 6 PM | Santa Monica, CA

An intimate, off-the-record dinner for senior media publishing leaders and executives to discuss what's actually working in audience monetization and growth right now.

This isn't a presentation—it's facilitated group conversations where you'll:

  • Share your single biggest challenge with peers facing similar issues

  • Participate in strategic discussions with executive-level professionals

  • Walk away with a blueprint not available publicly

  • Continue conversations in a supportive WhatsApp group

Location: Rustic Canyon Restaurant, 1119 Wilshire Blvd, Santa Monica

The dinner is designed to be high-trust, collaborative, and off-the-record. Includes dinner, drinks, and post-evening strategic discussions.

.NEWS OF THE WEEK.

➡️ AI-authored content just became normalized at another major publisher. Business Insider is preparing to launch an "AI author" byline for stories drafted by generative AI and edited by human staffers. The move follows Fortune's existing AI Intelligence byline and Bloomberg's AI summaries, showing accelerating publisher adoption. Axel Springer CEO Mathias Döpfner has mandated all employees use AI tools in some capacity. The test hasn't published stories yet, but marks another major publisher normalizing AI-authored content. Over the summer, BI editors told staffers they can use ChatGPT to write first drafts.

➡️ A test post crawlable only by Google appeared in Perplexity within hours. Reddit sued Perplexity and three data scrapers for allegedly bypassing access controls to obtain Reddit content at scale. The complaint includes this evidence of technical circumvention, along with data showing Perplexity's Reddit citations increased forty-fold after Reddit sent a cease-and-desist. Perplexity says it summarizes with citations and doesn't train on Reddit posts, but the technical claims echo Forbes and Wired accusations about undisclosed crawlers ignoring robots.txt. The case could set precedent for how AI assistants access and cite forum content.

➡️ Nearly half of AI-generated news answers contain significant errors. An EBU/BBC study of 2,709 responses across ChatGPT, Copilot, Gemini, and Perplexity found 45% misrepresented news content. Sourcing was the biggest weakness (31% with significant issues), with citations often missing or misattributed. Gemini showed the most problems (76% with significant issues), while others ranged around 37%. The study tested 14 languages across 22 public media organizations – accuracy problems included outdated info and incorrect characterizations of events. The EBU published a News Integrity in AI Assistants Toolkit alongside the report.

➡️ OpenAI now curates personalized article digests without users asking. The feature, called ChatGPT Pulse, automatically surfaces articles from across the internet based on conversation history, marking a shift from algorithmic feeds to AI-curated content bundles. CreatorIQ reports 1 in 10 agencies now work with Substack writers for content distribution, while the Financial Times questions whether we've "passed the peak social era" as audiences move to private networks. For publishers, the implication is dual optimization: video for human engagement, text and transcripts for AI discovery systems that train on earned media and blog content.

.SODP POSTS.

Featured Review

AlphaMetricX Review for 2025

Most media monitoring tools measure what's said about your brand but can't explain what it means or what to do next. AlphaMetricX positions itself as "media intelligence" – combining AI analysis with human expertise to transform data into strategic recommendations. Key differentiators: PR Impact Score weighs coverage by source credibility rather than treating all mentions equally, unlimited customizable dashboards without incremental costs, and AI-assisted query building. Strengths: comprehensive synthesis with strategic guidance. Limitations: custom yearly pricing lacks transparency, analyst dependency for smaller teams, limited integration documentation.

.JOB BOARD.

➡️ Business Insider needs lifestyle stories on travel mistakes, packing tips, lessons from retirees, and recipe reviews ($230+). Freelance opportunities for lifestyle writers with practical, service-oriented content experience. (Remote).

➡️ The San Francisco Chronicle needs full-time breaking news reporting interns, a food and wine reporting intern, and a summer data reporting intern ($19.18/hour). Multiple internship opportunities across news, features, and data journalism. (San Francisco).

➡️ Mongabay (Asia) needs a full-time wire reporter and a full-time engagement editor. Environmental journalism roles focused on Asia-Pacific coverage and audience development. (Asia-based).

.SOCIAL MEDIA.

➡️ Tim Ashelford on LinkedIn:

"AI can imitate [human creativity] – sometimes passably – but at its core it cannot have any value... A good story, like good wine, is valuable because of the unique circumstances that birthed it."

Ashelford, Managing Director at We Are Explorers (B Corp), published his company's full AI policy publicly—rare when most publishers keep policies internal. The outdoor publication's stance: no AI-generated sentences, paragraphs, or images on their site. AI permitted only for proofreading, research, and light editing.

The policy contrasts sharply with Business Insider's new AI author byline and Axel Springer's mandate that all employees use AI tools. As major publishers normalize AI-generated articles, smaller publishers like We Are Explorers are making anti-AI policies a brand differentiator. The strategic question: as AI content scales across major publishers, does editorial transparency become a competitive advantage or a capacity constraint?

These are the highlights for the last week.

Until next!

Vahe Arabian and the editorial team at SODP