You’re getting this email as a subscriber to the BrXnd Dispatch, a (roughly) weekly email at the intersection of brands and AI.
My friends Rich Ziade and Paul Ford, formerly of Postlight, now of Aboard, were kind enough to have me on their Podcast Reqless. I reached out after reading the introduction to their new podcast:
Why the name “Reqless”? Well—we’re not sure that AI is going to replace engineers, or artists, or musicians. It just doesn’t have the juice. But we’re increasingly sure that AI is going to replace…consultants. That doesn’t mean consultants go away. Just try to get rid of them! They’re everywhere. But it does mean the era of giant requirements documents may be coming to an end. Why spend all your time making a list of what to build when you could build it?
That obviously spoke to me and all the work I’ve been doing over the last few years. We caught up, clearly had a lot to say to each other, and decided to put it on tape (if you will).
One thing I discussed that I’m not sure I’ve discussed here is this pattern I’ve seen in all the big AI projects I’ve done lately with brands: The majority of the value happens at the intersection of AI (the models), code (for workflow), and human expertise (from the brand side).
At first glance, this isn’t particularly controversial: of course, good stuff exists at this intersection. I wrote a bit about this in my Creativity vs AI piece from June:
What I've seen in my own work collaborating with creative experts of all ilk is that when you take someone who is both talented and self-aware of the process to produce great work, it isn't overly challenging to design a system that can produce very high-quality output. But sometimes, that system requires a winding path. In one project, I encountered an issue where the AI was being too literal in its interpretations, producing output that wasn't quite right for the intended audience. The solution wasn't straightforward—simply adjusting the prompt wasn't enough. Ultimately, I found that introducing a second AI to review and refine the first AI's output was the most effective approach. This “AI reviewer” acted as a filter, adding a layer of interpretation that helped align the final result more closely with the project's goals.
Where the controversy lies is that the vast majority of AI tools out there are actually missing the expertise part. Don’t get me wrong, many of them, particularly the foundation models, are absolutely amazing. But they’re very reliant on your capability in an area to deliver above-average output. One of the issues I have with the majority of the AI marketing software out there is that they are missing the expertise bit—which isn’t really their fault since that lives inside the brands and is built on years of experience working with and building for the brand.
Another way to look at the triangle is as a matrix where you can clearly see what happens when one of the three components is missing:
AI + Expertise gives you capability amplification. That’s most of the good stuff folks are seeing with AI now. It’s what coders talk about when they describe how much they can get done using AI. They know the questions to ask and have deep expertise already, and AI is wildly accelerative with features like code completion, explanation, test writing, and the like.
Code + Expertise gives you regular software. This is the Sisyphean Hill we’ve all been climbing for the last thirty years. Software is built by encoding workflows (expertise), but it’s ultimately limited in what it can do because it requires that everything be deterministic and structured.
Code + AI gives you disappointment. Most of what we use when we use AI is actually code + AI. The code allows us to chat with it or build multi-step processes or kick out ready-made ads. The problem is that without expertise, you get the reaction we have all seen and heard where people ask the AI for something and then say, “That’s not very good.” Again, this makes sense: the baseline of these models can be conceptualized as a median of all the knowledge they’ve consumed. The output will also be average if you don’t layer on additional expertise.
Code + AI + Expertise will blow your mind. I’m legitimately unsure how else to describe it because it’s not really something we’ve ever seen. It allows you to take strategy and other special/unique knowledge and encapsulate it in a system in previously unimaginable ways. Whereas software in the past was limited to the data that was structured in a way it could fit into the database, the magic of AI is that it can reshape any data to fit into the structure you need. When you combine it with code, you can work through multi-step processes with ease—and if you think about anything interesting you do, it requires a multi-step process.
As we navigate this new world of AI, it’s becoming increasingly clear that the real magic doesn’t lie in AI alone or even in the combination of AI and code. The most interesting output comes when we bring deep human expertise into the equation. This trinity of AI, code, and expertise is where I’ve seen the most impressive results.
So, what’s the takeaway? If you’re a brand looking to leverage AI, don’t just shop for tools. Look for ways to infuse your unique expertise into the AI+code equation. And if you’re developing AI tools for marketing? Figure out how to tap into and amplify that brand-side expertise. That’s where the real value lies.
I think that’s it for this week. As always, thanks for reading, subscribing, and supporting. If you have questions or want to chat, or if you’re at a brand and try some of this out, please be in touch.
Thanks,
Noah