> ## Content Index
> Fetch the complete content index at: https://www.kendallmonk.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Brand New World
- URL: https://www.kendallmonk.com/brand-architect-its-not-personal/
- Published: 2026-11-07T03:27:51.000Z
- Updated: 2026-11-09T01:54:11.000Z
- Description: Feed AI a generic brand and it hands you back the average of lost. On creating Brand Architect, codifying human judgement, and using AI to put more humanity into the world, not less.
- Author: Kendall Monk
- Tags: Genesis, AI, Branding, Business, Authenticity, Identity, #genesis, #essay

**When we hand our brands to AI, do we hand over our humanity with them?**

---

*Written with AI, not by it.*

## If you can't beat it

The writing was on the wall. What we charged clients thousands of dollars for was about to be available for next to nothing. If AI was going to kill my business, then I had to make my business AI.

Not overnight, though. I couldn't drop everything and start building the AI version of a branding business. Still, I knew we needed to go all in. For many in business that means a focus on productivity: automate the admin, speed up the grunt work. But I was already seeing these models do something more interesting. They were super helpful in building brand profiles (and, to be fair, pretty bloody good ones), especially when you gave them high-quality context.

A good brand profile is challenging to get right for a bunch of reasons. It's nuanced and takes a lot of time, often with multiple hard to schedule interviews and calls. In person was the only real way to extract half-decent answers, to get inside the other person's head and find what we needed. Questionnaires never cut the mustard. No matter how open we made the questions, the answers were usually the bare minimum, not even scratching the surface. No one was there to push them harder, further, deeper.

After a little tinkering, it became clear an AI trained on my experience could do this as well as I could, in some cases better. It would come at clients from left-field angles, change tack, reword questions and suggest analogies that I would have struggled to imagine. It would keep going, strategically, comprehensively and efficiently, until it had what it needed, at the standard I’d set.

## Not safe for public consumption

I built a crew of AI interviewers, each drawing out part of a client's brand. We fed them everything we'd always done in analogue: the questionnaires, the interviews, our tricks that got a business owner to say what they actually meant.

These models are great conversationalists. That's what they're built for. But without clear boundaries they quickly become the neighbour banging on about a dog that died years ago when you're trying to get inside to feed your own. They just keep saying things, never knowing when to stop or that you actually have a destination.

Getting them to behave like the minions they needed to be, to do exactly what I needed every time without being told word for word what to say, took a year, on and off. Suffice to say, there were a shitload of versions. Loads of them were shit.

Give an AI interviewer a script and they'll just parrot the questions. Give them free rein and the quality is a shot in the dark. It's like training a working dog. Keep it on a lead and it'll never move a single sheep. Let it off too early and it's chasing roos to the next paddock. You want one that knows the job well enough to run wide and still bring the mob home. My Kelpie is good on sticks, not sheep. But then, I'm better with words than livestock. It takes two to three years to teach a dog how to do that. So, in theory, AI is easier to train. Like a puppy though, it can also piss all over your floor.

In-house, the tool was awesome. When an interviewer fumbled a question or wandered off somewhere odd, I had a hand on the tiller to keep the ship on course. Letting it loose was something else. It was not safe for public consumption. Without me at the helm, the risk of either a sloppy result or someone accusing me of robbery was high. So, back to the drawing board. Test. Sharpen. Repeat.

My architectural mind came in handy, but I'd never built anything like this: a system that worked as a loop, components that built on each other, got versioned, tracked and improved, and each time broke in new and interesting ways. Kaizen was my life.

It's not perfect, it probably never will be. But with each version, the gap between what the machine does and what I might have done gets a little narrower. That said, I'm not even sure closing the gap entirely is the best outcome.

## Strictly business

For a while the crew had human names (one was even called Miles). Now they just go by what they do, with a little whimsy for flavour: the Master of Messaging, the Purpose Wizard, the Voice Alchemist and the rest, each with a cartoon avatar to match.

Cartoons or not, the avatars are there for good reason. As are their personalities. People open up to something that looks and talks like a person. The more comfortable they are, the more honest the answers, and honest answers are what we are gunning for here. A brand profile built on mediocre answers is a mediocre brand. It's the questionnaire in a cool costume.

For a pack of minions, they're fun to chat to. Somewhere in a good session, the distinction between an AI and a human gets fuzzy. Nobody's deceiving anyone. The flags are all up: the names, the avatars, the fact you signed up to be interviewed by one. It's not a person and never pretends to be. In the flow of a good conversation, though, that fact ceases to be front of mind.

That fuzzy no man's land of identity is where a lot of the value of AI is, a conversation natural enough that you forget what's on the other end of it. There is also danger there. The same comfort that gets a business owner to open up could get someone to take advice they shouldn't, from something that sounds like it knows what's best.

In the scheme of things, Brand Architect is relatively low consequence. It isn't offering advice. It mostly extracts. It asks, it listens, it refines, and it writes down what it heard. It's not personal, Sonny. It's strictly business.

## Human software

Leaning in rather than trying to outsurvive it only works if you've got weight to lean with. A content production company I started in 2016 morphed into brand strategy after I came upon client after client who had no idea what they were saying, who they were saying it to, what they stood for and why. My approach had been tuned in the field and was already out there in one form or another. Brand Architect has made it accessible, interactive and intelligent.

Out of the box, the mainstream models are good, and can get you a fair way towards a functional brand profile. Getting it from functional (read generic) to genuinely yours is the gap I'm interested in – like going from flat-pack to hand-crafted furniture. It's essentially my years as a marketer, expanded into AI form: asking the questions I'd ask and the ones I'd never think of, pushing where I'd push and further, running every answer through the frameworks and the branding philosophy I'd honed. The humanity is what turns a good profile into a great one.

It also has advantages I could never afford. Time, for one. It isn't squeezing a client in between others, or running a business, or raising children, or sleeping. It's got one job and can do it whenever it works for the user. That's a massive plus for the end result. The user does the work when it suits them, when they've got the headspace and bandwidth to do a proper job, not when a client's diary momentarily lines up with mine.

There are two lots of humanity in every profile it builds: mine, codified, and the user's, drawn out of them one question at a time, at least as far as it shows up in their business. Their voice, values, purpose, insights and position. Mine goes in so theirs can come out. Call it human software.

## A proper job

Doing a proper job where it counts is crucial when we are handing so much over to AI. If we skim over the surface where we should be going deep, if we hand over the reins, then we will slowly homogenise our humanity. An average, an approximation. A copy of a copy of a copy, until it's just a smudge. The irony is, when used well, AI can help us inject more humanity into the world, not take it away.

There's a loop at work here, and it feeds itself. My instructions go into Brand Architect, the frameworks and judgement written down as honestly as possible. The questions come out. The user's answers go back in. The brand profile comes out. The profile then goes in as context for the business and whatever AI it runs on, and out come the strategy, the copy, the content. Feed it rubbish and it hands you rubbish back, then builds on the rubbish, until the whole business is a highly automated pile of shit. If we take the time, put in the work to push good material in, to consider our instructions and responses, the loop stays a positive one and the homogenisation (or worse, the system's bent) has less opportunity to take hold.

The usual advice is to pump in your website and your existing content and let the AI work out who you are, which is a bit like having your eulogy written from your LinkedIn. Feed it what a business has already said, though, and it'll say more of the same. Which is fine, if the business knew what it was saying in the first place. If not, it generates the average of lost.

A brand profile is the business's self, its foundation stone: who it is, what it stands for and why, in enough depth that anyone or anything working from it has more to go on than speculation. Plug that into AI and a business can build a lot more without losing itself. It stays consistent without being repetitive, because it has something that transcends the content itself to calibrate against. It stays authentic because it shares its DNA with the people who run the show.

## Eyes wide open

Brand Architect came before MYLS or Una. Looking back, it was my sandbox and proving ground, for capturing, reflecting and producing meaningful and authentic information, thoughts and stories, from and for human beings.

To get the interviewers right, I had to work out how to codify the way a person talks and behaves. In my mind I was going all in on AI marketing, but this was the beginning of me going all in, eyes wide fucking open, on AI – full stop. For better or worse, there is something about the convergence of human and machine operating systems that makes this work resonate with me more than anything else I have ever done.

I didn't need Brand Architect to make MYLS or Una. Neither is built on it. But after a year teaching machines to sound like people and to draw people out, it's hard not to wonder what else they could carry. A song that started as a letter to my dog. A prophet who happens to be a sloth. Or even a system that helps us think with depth and express ourselves with clarity. They all have their place and they all have their reasons. This brave new world is as wide as it is frightening. I'm trying to shine a little light on what's ahead, for my sake as much as anyone's.

When I saw the writing on the wall, I took steps to save my branding business. That path has led me further and deeper down the rabbit hole than I could possibly have imagined. I've been writing in some form my entire life, mostly in private. AI, though, I feel a very strong need to write about in the open. So am I now a Writer? And an AI Systems Designer? Both would sound cool, if job titles were still a thing anyone gave a shit about.

To meet the crew visit [brandarch.app](https://brandarch.app/?ref=kendallmonk.com&source=KendallMonk.com)

![Kendall Monk](https://storage.ghost.io/c/b4/37/b43713b8-6b65-4f47-9d83-1af2d6461741/content/images/2026/11/KM-postsig-dark-safe-thick.png) 

[Created with **mySignal**](https://mysignal.signalproject.co/?ref=kendallmonk.com&source=KendallMonk.com)

---