![]()
This post originally appeared on Untitled Thoughts on July 28, 2025 and has been adapted for Untitled AI. Read the post on SubStack here.
For this edition of Untitled Thoughts, we’re mixing it up. I’m handing the mic to my co-founder and Untitled’s VP of Marketing, Kramer Caswell. Kramer’s been deep in the trenches, thinking through how marketers succeed in a world where ecosystems shift their protocols toward agentic AI systems. This piece delivers a sharp take, challenging convention and offering a view into the future of mainstay channel marketing.
– Aaron Peabody, Co-Founder & CEO of Untitled
One of my favorite movies is Adam McKay’s The Big Short, the 2015 Oscar-nominated film about the housing crisis.
I could write an entire essay on it, but there’s one line that always stuck with me. As Ryan Gosling’s character breaks down mortgage-backed securities and subprime loans, he says the following:
”Wall Street loves to use confusing terms to make you think only they can do what they do. Or even better: for you to just leave them the f@!k alone.
Jared Vennett (Ryan Gosling)
I love that line, not only because it’s true, but because it reminds me of the marketing industry.
Marketers are pros at making things sound more complicated than they are.
We invent acronyms. We create entire disciplines. We rename the same thing six times and charge more for it every year.
And for decades, we’ve used that insider knowledge to feel powerful. Campaign setup. Match types. Negative keyword sculpting. Bid adjustments. Attribution windows.
All of it felt like a secret language that only we could speak.
Then along came AI.
At first, we loved it. We wrote 37 blogs with ChatGPT in one afternoon, updated every product page on our site. We turned a weekend into 3 months’ worth of SEO value.
But now that same AI is threatening to rip up the entire playbook we’ve been using since 2008.
LLMs like ChatGPT and Gemini are changing how people look for answers. If your customers stop Googling and instead ask an AI assistant, your #3 blog post about “best cookware for apartment kitchens” might not matter anymore.
There’s still value in creating content, but its role is changing fast. It’s less about keyword rankings and more about feeding algorithms training data, supporting brand credibility, or influencing what the AI surfaces in the first place.
However, artificial intelligence’s role in marketing won’t stop here. Rather, I believe we are entering a world that relies almost entirely on AI-agents or assistants for targeted marketing.
Untitled’s Prediction
Here’s my thesis:
Marketing channels in the next two years will get rid of almost all manual settings that you rely on today and your role will shift to “prompting” AI agents with the right data.
Every marketing platform you use, Google Ads, Meta, Klaviyo, you name it, is going to change. The knobs and dials you’ve spent years mastering? Gone. The strategies you optimized? Irrelevant.
You won’t manage campaigns in marketing channels.
You’ll “prompt” them.
What Will “Prompting” An AI Look Like in Digital Marketing?
When I say prompting, I’m not talking about a chatbox interface like ChatGPT. I don’t mean typing “build me a Facebook ad for camping chairs.”
I mean prompting in the form of injecting data and meta data. In other words, you are feeding marketing channels with audience insights like:
- First-party data: your customers, leads, subscribers
- Second-party data: like resolved website visitors or contextual intent data (learn more about second party data)
- Third-party data: data you’ve purchased or sourced
- And most importantly, conversion data: across the entire customer lifecycle
When we talk about conversion data, I’m not just talking about the purchase or “conversion” event you want to track. I’m talking about everything.
Everything that happens after the purchase.
Did they leave a review? Join your loyalty program? Start a referral? Expand their contract?
That’s the real prompt. That’s what the algorithm needs to find more customers that look just like your current ones.
An Example from Untitled’s Own Marketing
Okay, maybe you’re sitting here like, “Kramer, what the f@!k are you talking about? I can’t envision this.”
Let me explain it with Untitled’s marketing motions.
Some very quick background information: Untitled is a SaaS marketing product. We offer a free trial and primarily leverage Google Ads as a marketing channel and HubSpot as our CRM.
Here is the flow:
- Someone starts a free trial → we track that as a conversion event in Google Ads
- That user is created as a Marketing Qualified Lead in HubSpot
- When a user takes a specific action and/or replies to customer success or sales, they are marked as a Sales Qualified Lead and a deal is created in our pipeline with a contract value associated to it
- If they convert to a Closed Won deal → we push that deal value back to Google Ads so it knows what the user was worth, signaling to find us more people like this
While this is a simplistic view, many brands have run marketing like this for decades. The problem with this is that we are missing tons of conversion actions that, in an AI-driven world, our marketing channels need to know.
Our free-trial users and real customers are taking thousands of actions a day.
What if they activate key product features during the trial?
- That data sits in PostHog, our product analytics tool
- It doesn’t connect to HubSpot
- And it definitely doesn’t make it back to Google
So the most important actions, the ones that predict conversion, don’t get seen. That’s a problem.
If the AI doesn’t see the whole picture, it can’t learn what a high-value customer looks like.
This Isn’t a Theory. This Is Happening Now.
Let’s run it down:
- Google Ads deprecated Broad Match Modifier, updated phrase match to behave like broad, and is telling you to feed it audience + conversion data. Here’s the official help doc.
- Meta is actively building a system where all you do is upload your product catalog and budget, then it handles everything else: targeting, creative, bidding, optimization. By 2026. Yes, really.
- Adobe announced AI agents that will autonomously generate and optimize campaigns for brands like Coca-Cola. That’s already in market.
- HubSpot is launching its own AI agents to build campaigns and write emails.
- Salesforce, Klaviyo, Mailchimp, even TikTok. Every major player is moving in this direction.
We’re not waiting for the future. We’re in it.
So What’s Untitled’s Role In All of This?
We think of ourselves as the infrastructure between you and the machines.
You, the human with actual judgment, still need to make strategic decisions. But the execution layer? That’s getting abstracted away.
Our job is to make it dead simple for you to:
- Consolidate your audience data
- Track and score lifecycle behaviors
- Enrich that data with identifiers and intent
- And send it back to the channels that need it
We do all of this so these marketing channels can find more people who act like your best customers. Without the layer that sits in between YOU and the machine, digital targeting will be horrible.
What We’re Building Now
Untitled already helps brands:
- Sync first-party data
- Identify anonymous website visitors and returning customers
- Build audiences that go deeper than a pixel fire
- And run programmatic campaigns on Streaming TV channels
But what’s coming next is even bigger:
We’re launching a new Audience Builder that lets you push these enriched segments, with conversion values, to any ad platform you use.
Think of it as a universal prompt engine for paid media.
Your job? Feed it the right data. It’ll handle the rest (which sucks, but hey, we don’t get a say).
This Sucks For Digital Marketers, But There Is a Path Forward.
Let’s be honest, if you’ve spent the last 10 years mastering campaign structure and match-type philosophy, this shift feels awful.
But if you can adapt and learn to think like a data orchestrator instead of a bid manager?
You’ll be fine. Start preparing now:
- Combine your data sources
- Track the full customer journey
- Send lifecycle conversion events back to your platforms
- Think less like a technician, more like a systems thinker
And if you’re not ready to build all this yourself? Untitled is here for that too.
—
Join Us On SubStack: Untitled Thoughts
Untitled Thoughts is a space for my writing (and now, guests!) on marketing systems, identity infrastructure, data strategy, and life as a founder building in a messy, evolving landscape. I use it to share what I’m learning, what I’m seeing, and the ideas I believe more marketers should be paying attention to—whether tactical, philosophical, or somewhere in between. If anything here resonates, I’d love for you to test drive our platform at getuntitled.ai. Your feedback makes us better.




