My tune has changed over the past couple of years on how LLM-powered tools will shift the future of both the information ecosystem and human behavior. More in-depth thoughts in the future, but regarding AI as an workplace assistant: I've gone from "what do these bozos want me to do now?" to "this tech can be fun and helpful for me, even if it's used recklessly and badly by others."
Of course, there's much that needs to change in big tech and the AI industry, but if, like me, you are horrified by the violence of the present, it's comforting to consider how the future could play out more positively. I'm heads-down on big thinking.
Recently I've read several pieces on others' editorial processes involving Claude and other AI tools, and I'll share my own before long. That said, here's the disclaimer: I'm absolutely using Claude as an editorial assistant in my one-person newsletter shop. AI does not write for me ever, but it makes structural and logical suggestions that I wouldn't see otherwise. It helps with fact-checking and deep research. It probably tells me I'm a brilliant genius too often, but it's a helpful perspective to counter any internal monologue telling me I'm a failure who can't write.
And now, onto my experiences at ONA26.
–DC
Baby's umpteenth tradeshow: How the news publishing industry is surfing the AI tide
One of my favorite recurring magazine feature microgenres is the one where a writer attends their first-ever industry conference or tradeshow. The essays follow the tradition of New Journalism and David Foster Wallace, staggeringly first-person, as if the writer is the first to witness the absurdity of corporate showmanship. Here's one, another, and another.
These travelogues are meant to appeal to other freelance writers and are generally dismissive of the conference and tradeshow format, its tedium and brutalism. The authors scoff at the banality of chasing business goals so nakedly around folding tables in hotel ballrooms. The old reliable tricks of descriptive scene-setting bottom out at complaints about the aesthetic failings of functional corporate architecture. Inevitably, the first-person piece devolves into a frustration with the performance of capitalism— a diagnosis that is not incorrect, but misses the event's value by miles.
Business-to-business events have been a steady part of my career for the past sixteen years. I've experienced every angle of the conference business: I've worked for massive event companies, covered tradeshow floors to extract content, attended to learn, and led sessions and panels to share my work with an audience. I've been the buyer that booth staffers dream of, and I've been the marketer, ready to pitch passers-by. So I revel in these Baby's First Tradeshow essays, reading with immense condescension that mirrors the thinkpiece's assessment of the corporate culture scene. The wide-eyed writers, in dismissing the massive industry behind professional events, have failed to understand why business normies return to conferences year after year: at these events, the adults can make decisions that change culture without the burden of consumer gloss and inexperienced feedback. For most industries, there is exponentially more at stake on a tradeshow floor than there is in a year's worth of social media posts.
But conferences are also filled with normal, nice people doing their normal, nice jobs. No one's under the impression that they're at Davos. They're at the show because their boss sent them, presumably to learn or sell something. Maybe they're looking for a new gig, or their first gig, and most want to assess their place in a medium-consequential ecosystem. Others are on a normal business trip, seeking a few helpful tips or connections to bring back and make the investment worthwhile. The absurdity of the event's performance escapes no one, but it's a welcome break from daily routines. Might as well enjoy it.
Attending a conference is like watching a horror movie: for the whole shindig to make sense, it's best to suspend your disbelief. Yes, you will hear things that, if heard earlier in your career, would propel you to abscond to the woods and live off the grid. You will see the tackiest of slide decks and ingest heaps of cringe workplace humor. No matter what industry you're in, speaking corporate jargon is table stakes for participation, and if you're not fluent, the juiciest details will fly right past. Industry events require the radical self-acceptance of a character actor. My survival mantra: You are reading and responding to the room, which comprises very smart, lovely people whose company you enjoy. Just because you repeat a story does not mean you are becoming Brad Stand.
The Online News Association conference is better than most in this regard: journalists and their business counterparts value plain, direct language and won't hesitate to repeat your assertions back in their own words. Attendees run the gamut from newbie creators to executives of the most respected news organizations in the country. Everyone in the room cares deeply about journalism and their jobs, usually in that order.
On this excursion, I sought to take the pulse of how news publishing's fight for survival against both the dominance of big tech and the spectre of changing audience behavior. Sticking to sessions on the enterprise business and AI track, I wanted to assess how media companies are setting foundations for what's next.
The value of a conference rarely comes from information given during sessions or on the floor. It's in the opportunity for conversations among peers, in what is said and what's avoided, what's consensus and what's unknown. It's in listening for what's omitted, compared with other industries invested in similar topics. The tenor of conference conversation about how newsrooms are working now sets future agendas far more directly than the daily changes on a homepage, or posts from the digital rage monkeys.
Here's what I heard at ONA26:
At long last, multichannel audience acquisition is dominating over the easy wins of pageviews.
In a session on analytics from the nonprofit consultancy News Revenue Hub, most (but not all) attendees indicated that their higher-ups have moved on from the traffic era to a more diffuse view of audience acquisition. Speakers Sophie Ho and Abbey Gingras walked attendees through their methodology to understand what channels inspire audiences to subscribe and donate.
Based entirely on chatter and professional observation, nonprofit newsrooms are outpacing for-profit media companies in understanding the complex cycle of attracting humans to fund news. As an elder stateswoman of the analytics attribution era, I was pleased to see more advanced attribution modeling in place at the organizational level... because when I read mass media wonks and for-profit media execs on the same topic, I don't see the same level of sophistication in understanding why people choose to pay for news products.
News organizations are fervently experimenting with AI at the workplace level.
An early session by Reuters outlined their internal approach to experimentation with AI tools. While no employee is forced to use AI, they're certainly encouraged to experiment and build workplace tools—not for reporting, but for managing information, scheduling, and every other thankless workplace task that AI can support across organizations.
Is Reuters breathlessly using AI to understand and augment every last bit of the journalism business, replacing reporters with robots? Not at all. But they're building tools to administer AI access ethically and cautiously, taking a pragmatic approach to this new phase of digital transformation while enabling journalists to embrace at their own pace.
A later session with stakeholders from CNN, the Wall Street Journal, Reuters (again), and El Vocero de Puerto Rico covered similar ground. "Permission to fail" was repeated several times. Panelists mentioned motivating team members to experiment by celebrating AI "wins" internally, but none meaningfully addressed resistance from either employees or audiences.
Media companies are cautiously using AI tools to identify successful content and ideate more broadly.
As an analyst, I've always hated the leaderboard thinking of media company analytics. "Write more like today's best-performing story!" is rarely a recipe for success, and the neverending embrace of clickbait and trend piggybacking has certainly dampened innovation in news products over the past decade. But some AI-powered analytics approaches seem promising, especially when keeping a data expert in the loop.
New analytics tools, often engineered in-house, are finding opportunities for audience connection and brand differentiation through more in-depth insights. Instead of "do more like this story," the recommendation appears as "This topic performs consistently well, and no competitive organizations are covering it in the same level of detail. Do more like that." It's an evolution of the leaderboard, but it's better than pure vanity metrics. Since most local newsrooms don't have a trained analyst on staff, I'm hopeful these tools can identify more valuable data.
For-profit national media has sort-of gotten the hang of profitability on digital channels.
Compared with years of watching major media companies flail helplessly with every third-party algorithm update, it feels like mass media is taking ownership for product performance. A speaker from The Atlantic celebrated year-over-year revenue gains and new sponsorships (take what you can get?), and Crain's... honestly, I didn't hear what the guy from Crain's said. He was soft-spoken and drifting from the microphone. I was feeling a bit of FOMO from missing the concurrent session hosted the rockstars at WTF IS SEO?, who were cheering in celebration of their list hitting 14k subscribers.
Non-profit media companies are killing it in the face of garbage odds.
Nothing's better than a historic chandelier-speckled Grand Ballroom for a keynote. ONA's first-day anchor session featured leaders from Texas Public Radio, Public Media Corporation, Chicago Public Media, and National Public Radio. Although the U.S. pulled its federal funding from National Public Radio—17% of the total operating budget—none of NPR's member stations have shut down despite the funding pull. Their success persists due to collective, coordinated efforts and shared services. Plenty of people still listen to radio, both directly and through podcasts, and audiences donate even when they don't regularly listen. I'm an afar admirer of non-profit news—I could never hack it in the hyperpolitical world of public media—but their human-first model is one to follow. No one at the keynote even mentioned AI.
Between sessions, I spoke directly with every table that was not recruiting. I know better than to prod part-timers for details, but I still asked table staff specific questions about how their software products function. Sometimes, I received answers. I let vendor reps scan my badge, possibly the most adult transaction one can initiate outside of sex work or drafting a will. I am a buyer, the target audience, but it still feels a bit weird: Add me to your company's mailing list. I trust you not to spam me. Maybe one day I will buy a product.
Throughout the show, I listened earnestly and interacted awkwardly among the smart, lovely people. Conversation was functional and often humorous, but not thrilling or dangerous, which, thank goodness. Perhaps it was the tempering presence of public and nonprofit media, but no speakers sounded reckless or obtuse in the attention-seeking manner of media business podcast guests. I felt at home among hopeful normies who also harbor the Xennial belief that we can change the world with our work, even in the worst of circumstances.
What I listened for and didn't hear at sessions or on the floor:
No one mentioned backlash and online discourse around journalists' and writers' use of AI.
I'm seasoned enough as a consultant to accept that executives are not having the same conversations about technology that their workers are having. However, in the case of newsroom AI, the giant knot of mixed signals requires untangling. In the case of for-profit media companies I'm nicely suggesting a little more rigor and candor when it comes to thorny ethical and legal issues around artificial intelligence. I've been following AI developments as part of my consulting practice since 2019, and I'm confused, so I can only imagine that a career sports or culture reporter must feel something like whiplash.
Yes, newsrooms want to prepare their talent for the future, and AI-powered tools can support journalists at work. But the same media companies are suing AI companies for copyright infringement while they are encouraging workers to experiment with LLM products from the same companies they are suing. Meanwhile, online, a socially influential community of writers and workers are vehemently rejecting any and all acceptance of artificial intelligence integrated in writing work processes. If you're at a media company where higher-ups are prodding you to experiment but the internet is shaming you for participating, you're right to be confused! Where does "permission to fail" start and end in newsroom AI experiments, given the broader conversation?
Audience information-seeking behavior is changing, rapidly.
People are using LLMs as an augmentation of or replacement for traditional search engines. Younger audiences are heavily relying on chatbots for all information access. If every workplace in the United States weren't experimenting with AI, behavioral change would move far more slowly. But as adults are prodded to find a use for AI in the workplace, many are succeeding and, like me, finding begrudgingly that LLMs can be pretty great for organizing work and finding information. We're even somewhat excited about it.
In ten years' time, interaction with news organizations will look very different. Natural language interaction will replace web search eventually, but aside from a couple of mentions of "Google Zero," neither presenters nor vendors outlined how publishers will ride the wave of the sea change.
What are the technical solutions for maintaining factual accuracy among token prediction machines?
I heard talk of data cleanliness and garbage-in-garbage-out, but other than that, structured data was not a topic of conversation. The one person to bring up knowledge graphs was me. I guess I'll have to wait for Knowledge Graph Conference next month, and I really need to start studying or else it will be way over my head. One day editors and publishers will be in charge of maintaining knowledge graphs, but not today.
Are media companies upgrading ad tech so that the audience experience is better?
For an industry that has spent a decade railing against Google and Meta, I heard very few solutions to the absolute garbage experience of ad tech. Publishers could be building innovative solutions to solve for the fact that ad tech hurts their product perception. And ad tech could be easily improved! Publishers could be assessing whether AI-powered revenue generation tools are consumer-friendly and trustworthy. Are AI-powered dynamically priced paywalls another word for price-gouging? Is there a way to build a better, more contextually relevant ad tech product, so the industry no longer siphons away a third of its revenue to third-party software? Can we build ads so that smart audiences are not hit in the face with dongs as we are trying to read the evening news?
What are other design-first product improvements publishers can make to attract audiences and keep them engaged with news?
Design is a differentiator. I suppose I will have to wait for the News Product Alliance conference in October to hear more about on-site search, interactivity, and other product upgrades that are more in my wheelhouse.
These conversations will be had elsewhere, or not at all, as the information business continues to fragment. In the meantime, I had a lovely time meeting some internet acquaintances in real life and meandering about Chicago, a top-tier American city. I did not leave feeling disgusted with capitalism, at least not more than any normal day. I did not become Brad Stand, but I did get what I wanted: a pulse. At conferences, industries reveal what normal looks like for their business. News is operating as it always has been, for better or for worse. Next time I will bring more solutions and work to share, or at least some stickers.
Content tech links of the week
Back with links! I have been trolling around on the trash heap of the Substack app, and I really wish the damn thing had a bookmarking feature. It's an interesting ecosystem, mostly amateur, but I'm gradually locating hidden gems. I'm also on the lookout for good business newsletters in general, so send me your recommendations or your own writing!
The below may actually be the links of last week, or even last month, but they are good nonetheless:
- How agencies distort the perception of fandom, by musician Eliza McLamb. Because I've worked on online culture for too long, I know that nothing "organic" on the internet is actually organic. Popular accounts buy followers or use engagement hacks. Good websites pay folks like me to structure their content for discovery in search engines and elsewhere. And, given what I know about music industry marketing, the Narrative Strategy agency work that McLamb outlines here doesn't surprise me... but it's still frustrating.
- The discourse on journalism, AI, and writing, contextualized by Max Read. I'm going to love any essay that starts with a picture of the Mechanical Turk, and Read summarizes the aforementioned Online Writer AI Discourse evenhandedly.
- What does it mean to be "solutions-oriented"? Jessica Talisman generally sticks to detailed descriptions of knowledge graph infrastructure—which I am studying up on before I see her at next month's Knowledge Graph Conference—but this accessible essay on the concept of Solutionism crosses over into a less niche domain.
- Language models as a bureaucratic technology. I'll admit that lots of academic writing about AI and language flies right past the business context I seek out, but Ben Recht's writing sits close enough to practical application that I can see how academia fits in the puzzle.
The Content Technologist is a newsletter and consultancy based in Minneapolis, working with clients and collaborators around the world. It is written and edited by Deborah Carver (me), an independent content strategy consultant. Claude augments the editorial process and helps with fact-checking and writing clarity, but all words and ideas are my own.
Affiliate referrals: Ghost publishing system
Cultural recommendations / personal social: Spotify | Instagram | Letterboxd | PI.FYI
Did you read? is the assorted content at the very bottom of the email. Cultural recommendations, off-kilter thoughts, and quotes from foundational works of media theory we first read in college—all fair game for this section.
Between the new Sturgill Simpson / Johnny Blue Skies and the new Robyn records, I'm glad my generation still has room for fun-lovin', lovin'-lovin' danceable, dopey bops.
