Building with AI in Large Successful Companies // BRXND Dispatch vol. 71
On Amazon, Adam Driver, and AI ads
You’re getting this email as a subscriber to the BRXND Dispatch, a newsletter at the intersection of marketing and AI.
Luke here. With Noah on vacation, I’ll be holding down the fort. Today I have two things to share: First, a recap of an amazing story from our recent LA event about how Amazon first started dabbling with AI to create ads, culminating in a series of major spots featuring Adam Driver performing real customer stories. Second, I’ll share some thoughts about what Deep Research means for content creators and intellectual property.
Before getting down to it, I want to make sure everyone saw the news that BRXND NYC is returning on September 18, 2025.
General tickets aren’t on sale yet, but you can leave your email to be informed when they are.
If you’ve never been to a BRXND event, the basic approach is covered by the programming tenets:
Doers > Speculators: We want BrXnd to be about doing and learning, not talking and speculating. We are too early in this journey, and the technology is too different from what we’ve seen before to get stuck speculating about its future impact.
Curiosity > Overconfidence: We embrace the unintuitive nature of AI and its future implications. We're all learners in this new era of technology.
Conversations > Panels: Whenever possible, it’s better to have just two people on stage talking to one another rather than a whole set of speakers competing for time and attention. Go deeper with fewer people so that the audience can learn from their thinking and experience more than just their opinions.
Show > Tell: It’s always better to show people something than to tell them about it. While other conferences try to urge people away from showing specific products, we think it’s most useful for people to see what’s real and possible.
Process > Perfection: Take your audience on a journey, even if it’s something that didn’t work out. This is an audience of doers and experimenters, and understanding the journey is often more important than just seeing the outcome.
The plan is to keep it pretty small (around 200) and make sure it’s full of interesting people and featuring talks, demos, and presentations focused on what’s possible today instead of what might happen in five years.
We’ll have lots more info coming soon, including how to get tickets and who will be speaking. Of course, if you’re interested in sponsoring the event, we’d love to talk. If you have speaking ideas, we’re happy to hear those as well (though please make sure to read the tenets and explain how your proposal fits).
Making AI Ads That Don’t Feel AI-Generated
Last week I wrote about Noah’s opening speech at BRXND LA and how it provided a roadmap for marketers who are butting up against restrictive bureaucracy in their quest to explore AI.
Noah’s talk was the perfect setup for a presentation later in the day by creative director Mike Houston of Amazon about how to actually get AI projects off the ground in a large organization. (I strongly recommend watching the full presentation below.) What struck me most wasn’t just the specific examples of how Amazon has used AI tools to make ads that don’t feel AI-generated (though those were great), but the underlying strategy to breaking through organizational barriers that often seem insurmountable.
A few key takeaways for anyone trying to push AI initiatives forward in their organization:
Work with legal early and treat constraints as parameters rather than blockers
Create regular forums for sharing experiments and learning (the cadence creates momentum)
Look for ways AI can enhance existing processes rather than replace them entirely
Don’t be afraid to show the failures - it sets realistic expectations and builds trust
Focus on finding and amplifying human stories rather than generating artificial ones
My favorite part was Amazon’s approach to customer stories. Instead of generating fake AI content, they used automation to parse through Amazon’s massive review database, looking for compelling human stories. They even created a scoring system to surface reviews with great metaphors or descriptive language. The end result was a series of ads featuring real customer stories performed by Adam Driver - authentic content that they found through clever use of AI. It’s a perfect example of how AI should work: not replacing human stories, but helping us find the best ones hiding in plain sight.
If you missed any part of BRXND LA, don’t fret; all the talks will be on YouTube soon.
What Else Caught My Eye This Week
There’s an emerging narrative in AI circles that research assistants like Deep Research are about to trigger a massive shift in how organizations think about their intellectual property. The conventional wisdom suggests we’re headed toward an era where high-value data becomes increasingly difficult for LLMs to scrape.
Ben Thompson made this point recently:
There is a good chance that Deep Research, particularly as it evolves, will become the most effective search engine there has ever been; it will find whatever information there is to find about a particular topic and present it in a relevant way. It is the death, in other words, of security through obscurity.
Unless, of course, the information that matters is not on the Internet.
I believe this line of thinking is generally correct. As we recently noted on our Alephic blog, companies should think carefully about letting public models train on their “private tokens,” or the unique knowledge, processes, and language that give them a competitive edge. And, in the wake of Deep Research, we’ll likely see a rush to privacy from organizations where proprietary knowledge is the core business: legacy publications protecting their journalism, hedge funds guarding their trading strategies, scientific institutions safeguarding their findings, and so on.
But there’s an interesting knock-on effect to consider, one that Noah pointed out recently on X: Who is actually incentivized to make their thinking public in this new media landscape?
As more paid content creators move their work behind paywalls, a clear counter-incentive emerges: those seeking influence—brands, think tanks, anyone with an agenda—will benefit from doing the opposite. What we could be looking at is a massive swing back to content marketing.
In a world where AI systems are becoming key mediators of information, having your content readily available for these models to learn from could be incredibly valuable. While conventional wisdom suggests the public web will devolve into AI-generated slop, the incentives actually point toward investing in high-quality, open-access content that can shape how the next generation of AI systems understand and represent your domain.
That’s it for now. Thanks for reading, subscribing, and supporting. As always, if you have questions or want to chat, please be in touch.
Thanks,
Luke