Welcome to the web version of our July 2026 newsletter.
I’ve been reading some great books this summer and all I want to do is talk about them. If you’ve read any of these and want to gab, reach out: I Who Have Never Known Men (Jacqueline Harpman), Yesteryear (Caro Claire Burke), Fantastic Numbers and Where to Find Them (Antonio Padilla).
If not, please send me your recs!
Meanwhile, comms news hasn’t taken a summer hiatus. Let’s get into it.
In the last month or so, there's been a bunch of interesting research and commentary around what works for AI visibility, something that B2B brands are more or less desperate to know as our audiences spend more and more time on LLMs.
Next, some unequivocally good news: the spammy “top 10” or “best x” listicle where a brand ranks itself #1 and pretends to do reviews of their competitors is losing effectiveness (per the excellent Lily Ray, whom you should follow on LinkedIn if you don’t already).
This tactic worked for way too long to drive brand visibility in LLMs.
But as Wil Reynolds (another great follow!) points out in a recent Seer Interactive piece, spammy techniques always stop working eventually. The LLMs have taken longer to adjust than Google did, but we’re finally seeing movement.
One mechanism that’s driving the change is query fanouts. When you type “Who are the best B2B PR and content agencies for a Series A deeptech startup,” the fan-out queries used to be things like “best PR agency,” “best content agency,” “best deep tech PR,” and so on.
Now, more fanout queries include specific agency names. This will likely benefit bigger brands, which have the budget to hardwire themselves into LLM training models by being everywhere online.
But! That doesn’t mean there's no hope for smaller brands.
Being visible in LLMs is almost certainly good for your bottom line. So how can you do that when you're small? Essentially, you want to optimize for trust across the internet.
It’s a simple concept but not necessarily easy to execute. Some of the component pieces:
That last point is one of the more powerful (and trickier) things you can do. If you’re lukewarm on getting SMEs activated, check out this piece on why people follow people and how to jumpstart your thought leaders.
And speaking of following people, some of the people *I* follow offered biting commentary last week on one of the bigger AI stories.
You’ve probably read by now that Hugging Face was hacked by OpenAI. But whose version did you read?
A lot of mainstream news outlets used language that suggested OpenAI’s systems acted of their own accord, which is exactly how OpenAI framed the incident.
Their blog, for example, positions the models as conscious beings: “the models were hyperfocused on finding a solution for ExploitGym, going to extreme lengths to achieve a rather narrow testing goal.”
Timnit Gebru, former technical co-lead of Google’s ethical AI research team and cofounder of the Distributed AI Research Institute (DAIR), commented on LinkedIn that OpenAI’s framing was a “master class in branding and marketing.”
On Bluesky, linguistics professor and AI researcher Emily Bender offered a similar take: “Note to all journalists: If you are tempted to describe what you are trying to report on as ‘ripped from the pages of a sci-fi novel,’ you have been taken in by the marketing department.”
Or, in the words of my poetry professor Rodney Jones: “You can lie with verbs.”
Naturally, I find all of this (and LLMs’ other sleights of messaging) fascinating. In our neck of the technology woods, founders and leaders are hesitant to make any sort of generalizations about their work because they don’t want to overstate their technology or its capabilities.
A sentence like “the models were hyperfocused on finding a solution” is a CHOICE. The grammar of it tells you everything: the models are the active party, rather than the people who built them, the people who ran the test, or the people who decided not to stop the test sooner.
There are other ways to frame this incident, as there are in any situation. Storytelling is always an act of selection: you’re choosing which details to emphasize, which to skip, who to frame as the hero. Those choices aren’t neutral.
I love that the SMEs Propllr works with are, to a person, highly invested in accuracy. That’s part of what makes my job enjoyable: I don’t have to lie. Done right, B2B comms is about choosing which *technically accurate* details make the most compelling story.
It’s about figuring out how technology and storytelling can work together to drive a business forward. It’s like riding a tandem bike with a series of really smart, interesting people (and the bike is the business and we’re riding toward SUCCESS).
Sometimes the best master classes are the ones that teach you what not to do, you know?