The AI Assumptions That Hurt Your Brand

Early in my career, I made a bet that a lot of founders thought was backwards: I'd rather hire a marketer before I hire a salesperson.

I was fortunate to work with incredible marketing leaders and I saw firsthand how these leaders created a flywheel that accelerated sales. Leads show up already warm, already curious, already halfway convinced. The sales team stops chasing and starts closing. I've watched this play out enough times now to hold it as one of my foundational truths: marketing is one of the most powerful levers a company has for growing revenue, because it's usually the first interface someone has with your brand. Done well, it's not just informational. It's magical. It makes someone feel something, and that feeling is what gets them to raise their hand, sign up, or pull out their wallet.

And yet marketing is a strange department to defend, because it's asked to be everything at once. You're the team that "makes things pretty." You're also the team watching ad spend down to the penny. You're the team that drops what it's doing when product needs a data sheet by Friday or sales needs a presentation tonight. I'd argue there's no other function in a company more tied to its reputation, and also no other function held to a more subjective standard. Everyone has an opinion on what looks good, what sounds right, what feels "on brand." I’ve personally wanted to tear my hair out trying to please a room full of stakeholders on messaging.

Now let's sprinkle in AI

Generating content, images, and even entire websites quickly is table stakes now. A single prompt gets you a blog post, a hero image, a landing page, before your coffee gets cold. Speed is useful, but it's not giving you any competitive edge. If you can generate an asset with a one shot prompt, so can your competitor. Marketing leaders should ask an honest question: if this is the bar, what is it we actually add?

Here's my answer, and it's the opposite of doom and gloom: the function of marketing just became more important than it's ever been.

💡AI chatbots are the new first stop before your homepage, and you have zero visibility into how they're describing your brand.

AI is now the interpretation layer between your brand and your audience. People are asking ChatGPT what a company does before they ever hit your homepage. And we don't control that layer. We don't see inside the model to know what it's gathering about us, how it's characterizing our product, or why it recommended a competitor instead of us. The signals we used to obsess over, keyword rankings, SEO position, are being replaced by something closer to a black box. We're navigating blind, hunting for any signal we can actually measure.

This is good news (even if it doesn't feel like it yet)

I mean that. For years, marketing rewarded whoever could write the best clickbait or outspend everyone else to own a keyword. That's not where the leverage is anymore. What's left, once those tactics stop working, is the actual craft: understanding who genuinely matters to your business and being strategic about reaching them. That's the version of marketing I fell in love with in the first place. It's a good time to be someone who takes this seriously.

But the part that's not sustainable is that most teams have used AI to move faster, post more, produce more, without asking whether "more" was ever the goal. I talk to marketers constantly who've increased their output and their spend, and are getting flat or declining results for it. Same budget and less to show for it. The same playbooks aren’t yielding the results we’re used to seeing. All of this has resulted in more noise, faster.

“Undust” the furniture

And the best way I can explain why I think this is happening is from a children's book series from my childhood, Amelia Bedelia.

If you didn't grow up with the books by Peggy Parish, Amelia Bedelia is a housekeeper who works for the Rogers family. She desperately wants to do a good job, but takes every instruction literally. When Mrs. Rogers tells Amelia to "draw the drapes" at five o'clock, she sits down and sketches them. When Mr. Rogers tells her to "put out the lights," she unscrews the bulbs and sets them outside on the lawn. Eventually the Rogers learn to stop saying "dust the furniture" and start saying "undust the furniture," so Amelia Bedelia would do the task that they intended.

I think about those books all the time as we've been building OrcaPanda. Miss the context behind an instruction and you get a technically correct answer, but not the result you were looking for.

It’s a pretty common issue I see with marketing prompts and skills. When you tell a model "you're a world-class copywriter, write me a blog post about X," you're handing over an enormous amount of unstated assumption. What do you mean by world-class? Why this topic, right now? What is this post supposed to do for the business? None of that is in the prompt, so the model fills the gap with its best guess, the same way Amelia Bedelia filled the gap with hers. The output isn't wrong exactly. It's just an answer to a question you never asked. Do that enough times across enough campaigns and you're not building a brand, you're building noise, the exact kind of noise AI itself now has to sift through.

The real problem was never context. It's continuity.

Most marketing leaders I talk to have already figured out step one of the fix: get the AI more context. Load in the brand guide. Connect the tools. Write some shared prompts or skills so at least the output sounds more like you. That helps in the beginning, but it's not the actual solution.

Every team already has context. It's just scattered across brand guides, old decks, a hundred Slack threads, and whoever's been there long enough to remember why a decision got made. But context rots fast. Positioning changes. A feature launches. Pricing shifts. Someone joins or leaves the team. None of those places reconcile the change on their own, which is why a living document is not really living. It’s a document that sits in a folder and is revisited once a year.

This is the part that general AI doesn't solve on its own. These models are extraordinary, and they are still general by design. They don't know that your CEO hates the color pink for reasons he's never once explained (true story), that legal will kill a claim your last draft made, or that the last campaign underperformed because the pricing page confused people.

Context is a snapshot, but continuity is what carries it forward into the next initiative, decision, and team understanding.

That gap is what we built OrcaPanda to close

Most tools treat context like a static packet: upload the brand guide, paste in a prompt, connect a data source, and call it good. That's not what context actually is inside a marketing team. OrcaPanda carries what happened forward into what happens next.

Managing that constant flow of information across your team, documents, and meeting transcripts is difficult to do manually. I've felt this myself more times than I can count. OrcaPanda is built to do that work: figuring out what actually matters, telling the difference between something confirmed and something someone's just guessing at, keeping it current as the business changes, and pulling the right slice of it into whatever work is in front of you.

That's not abstract. OrcaPanda is catching that a piece of messaging you retired six months ago is still showing up in this week's assets, before it ships. It's noticing that a competitor's product page has started to look close enough to yours that the thing which used to set you apart is quietly blending in, before a customer has to be the one to point it out to you. It's flagging that the same pricing-page confusion that quietly sank a campaign last quarter is about to repeat itself in this week's launch, before the budget's already spent.

We've run more than 100 brands through OrcaPanda, and this continuity gap exists even among the strongest teams. Without a shared foundation in place first, AI doesn't scale a team's marketing. It multiplies the cleanup and teams ended up spending more time correcting, re-explaining, and re-prompting than they would have spent doing the work by hand.

💡AI doesn't eliminate ambiguity. It multiplies it across every channel before anyone notices.

Our philosophy

The evolution has gone from static software, to AI with an empty chat window, but OrcaPanda starts from the stance "given everything we've learned so far, what's the next best thing to do?"

To accomplish this, we built OrcaPanda so it would accumulate understanding alongside your marketing team, not just process what it was told. It adapts to how your team actually works instead of asking your team to adapt to it. Continuity is what makes that understanding possible in the first place.

OrcaPanda gives your marketing team a shared system where that understanding and your organization's accumulated experience work together. It learns your business and keeps that understanding current, not just against what's true inside your own walls, but against what's changing outside them: what a competitor just repositioned around, what an audience is responding to this quarter that they weren't last quarter. That knowledge doesn't sit frozen in a deck from two quarters ago or wait for someone to notice it's gone stale. It stays current because OrcaPanda is the one keeping it that way, continuously, not because someone remembered to. OrcaPanda unlocks a new persistent marketing capability.

We see this as the difference between software and a true extension of your team.

Marketing shouldn't reset after every campaign

Every launch should make the next one smarter, not because someone remembered to write a recap, but because the understanding behind the work carries forward on its own. That's your campaigns compounding, and it's the difference between a tool that helps you produce more and marketing that gets better every time it's used. A record of what happened and why. Now your team spends its time on judgment and strategy instead of re-explaining itself to a machine that forgets.

More output was never the goal. Anyone can generate assets, including your competitors. The hard part, the part that was always the real job for marketing, is knowing what's worth saying, to whom, why, to produce your intended outcome. That's not something AI took away from marketing. It's the thing AI just handed back to us.

If any of this named something you've felt but couldn't quite explain, I'd rather show you than tell you more. Let's start there: I'll walk you through what OrcaPanda discovers about your brand, and we'll begin building the foundation that lets your marketing scale.