← Back to Brain Food AI Ads Are Coming: Here's What I Learned Testing This Across My Businesses (And What You Need to Know) Mixed

AI Ads Are Coming: Here's What I Learned Testing This Across My Businesses (And What You Need to Know)

• 9 min read • lab-reports

AI ads are coming fast. I tested ChatGPT, Perplexity & Google's AI ads across my businesses - here's what actually works, what failed spectacularly, and your action plan.

Fair warning: this is going to be a long one. But stick with me - by the end, you'll understand the AI advertising landscape better than 90% of marketers out there.

Listen, I've been testing AI tools across Rocket Media and Digital Ignitor for the past 18 months, and here's the honest truth: the way we think about search and advertising is about to get flipped on its head.

Remember when Google Ads was the new kid on the block and everyone thought banner ads would rule forever? We're at that same inflection point right now with AI search ads - except this time, the shift is happening 10x faster.

I spent the last three weeks diving deep into this rabbit hole, running experiments, talking to other agency owners, and yes, failing spectacularly at a few attempts to "hack" the system. Here's what I learned testing this in the real world...

The Simple Truth About AI Search (That Nobody Wants to Admit)

Let me save you the trial and error I went through: AI search isn't just another channel - it's a complete rewiring of how people find information.

My 15-year-old just asked ChatGPT to help him pick soccer cleats yesterday. Not Google. ChatGPT. When I asked why, he said, "It actually answers my question instead of showing me ads."

Ouch.

But here's the kicker - those same AI platforms that promise "no ads" are all racing to add them. And after testing this across both my businesses, I can tell you exactly what's coming and how to prepare for it.

What's Actually Happening With AI Ads Right Now

ChatGPT: The $25 Billion Question

Here's what I learned after wasting weeks trying to get ahead of this: OpenAI is playing the long game, and it's brilliant.

They've been saying "no ads" publicly while quietly hiring advertising veterans from Google, Meta, and X. Internal documents suggest they're planning to generate $1 billion from "free user monetization" by 2026, growing to $25 billion by 2029.

Sam Altman went from calling ads a "last resort" to saying he's "not totally against it" on their recent podcast. That's corporate speak for "we're definitely doing this."

What this means for your business:

Perplexity: The Testing Ground

I've been running experiments with Perplexity's sponsored follow-up questions since they launched, and here's the framework that finally made this click for me:

It's not about interrupting - it's about continuing the conversation.

Their ads appear as sponsored questions like "How can I use Indeed to enhance my job search?" after answering your original query. The AI still generates the answer, but the advertiser pays for the question placement.

After testing this for one of our HVAC clients, here's what actually moved the needle:

Real failure I had: Tried to force product-focused questions. Complete disaster. The platform punished us with terrible placement. Switched to education-first questions ("Why do heat pumps fail in winter?") and suddenly we're getting quality leads.

Google: The Sleeping Giant Wakes Up

Google's already showing ads in AI Overviews and their new AI Mode. I tested this across 15 different client campaigns last month. Here's the simple version:

You can't opt out. And that's actually good news.

Every Performance Max, Shopping, and Search campaign using broad match is automatically eligible. No special setup needed. But here's what nobody's telling you - the results are wildly different from traditional search ads.

The Framework That Makes AI Ads Actually Make Sense

After failing at this spectacularly for three months, I finally discovered the pattern. Think of it like building an organized system for your business:

Traditional Search Ads = Fishing With a Net

AI Search Ads = Having a Conversation at a Party

Here's the framework I use now:

The 3-Layer AI Ad Strategy:

  1. Discovery Layer - Educational content that answers the "why" and "how" questions

  2. Consideration Layer - Comparative content that helps with decision-making

  3. Decision Layer - Specific solutions with clear next steps

This isn't as complicated as the experts make it sound. You're just matching where people are in their thinking process.

What Actually Works (Based on Real Testing)

1. Stop Thinking Keywords, Start Thinking Questions

I tested this across both businesses. Keywords are dead. Long live natural language.

What failed: "HVAC repair Phoenix" (our bread and butter Google Ad) What worked: "Why is my AC making weird noises at night?"

Users ask AI questions averaging 23 words. Our traditional Google searches? 4 words.

You're probably overthinking this - here's the simple version: Write ads like you're answering a friend's text, not bidding on an auction.

2. The Authority Signal is Everything

Here's what I learned testing content across 50+ client sites: AI trusts what others say about you more than what you say about yourself.

We had a client with perfect on-page SEO, gorgeous website, zero AI visibility. Added three Reddit discussions about their service, two third-party review mentions, and a local news feature - boom, showing up in ChatGPT responses within two weeks.

Your action plan:

3. The Technical Stuff That Actually Matters

Fair warning: this takes time to get right, but the payoff is worth it.

After testing 15 different technical approaches, only three actually moved the needle:

Must-Have Technical Optimizations:

I failed at this spectacularly when I tried to get fancy with dynamic content loading. Lost 3 weeks of AI visibility because ChatGPT couldn't crawl our pages. Sometimes the simple approach is the right approach.

The GEO Revolution (And Why You Should Care)

Everyone's throwing around "GEO" (Generative Engine Optimization) like it's the next big thing. After 6 months of testing, here's the honest truth:

GEO is just SEO with better manners.

Instead of stuffing keywords, you're:

The framework that finally made this click for me: Write like you're explaining to someone who's smart but busy.

What This Means for Your Business (The Real Talk)

Let's strip away the hype and focus on what actually matters for your business:

If You're in B2B:

You absolutely can do this. Focus on educational content and thought leadership. AI platforms favor expertise over advertising, so your deep industry knowledge is your superpower.

If You're in E-commerce:

Start optimizing product feeds NOW. Shopping ads in AI are all about comprehensive product data. Every missing attribute is a missed opportunity.

If You're Local/Service-Based:

This is your moment. AI search is AMAZING at local intent. "Find someone to fix my garage door today" is exactly what these platforms are built for.

If You're Not Sure:

Start with one piece of comprehensive content answering your customers' biggest question. Test it, refine it, track AI visibility weekly. You can do this in 5 hours, not 5 months.

My Predictions (Based on What I'm Seeing)

After testing this across multiple businesses and watching the patterns, here's where I think we're headed:

By End of 2025:

By 2026:

By 2027:

Your Next Steps (The Practical Stuff)

You're probably overthinking this - here's the simple version of what to do next:

This Week:

  1. Search for your business on ChatGPT, Perplexity, and Gemini

  2. Note who shows up instead of you

  3. Find one place they're mentioned that you're not

This Month:

  1. Create one comprehensive piece of content using the GEO framework

  2. Get it mentioned in 3 third-party sources

  3. Add FAQ schema to your top 5 pages

This Quarter:

  1. Test Perplexity ads if they're available in your market

  2. Ensure all Performance Max campaigns use broad match

  3. Set up tracking for AI referral traffic

The Honest Truth About What's Coming

I've made plenty of mistakes testing this stuff. Wasted money on the wrong approaches. Built systems that completely failed. But here's what I know for sure:

This shift is happening whether we're ready or not.

The businesses that adapt now - while there's still arbitrage and opportunity - will own the next decade. The ones that wait will be playing catch-up, just like the companies that ignored Google Ads in 2005.

You don't need to be perfect at this. You don't need a massive budget. You just need to start testing, learning, and adapting. The framework is simpler than you think, and you absolutely can do this.

One Final Thought

My 9-year-old asked me last week if AI was going to replace advertising. I told her no - it's just going to make it more helpful and less annoying.

That's the opportunity here. We get to be part of making advertising actually useful again. No more interrupting. No more tricking people into clicks. Just helpful answers that happen to come from brands.

Yes, there's a learning curve, but here's how to flatten it: Start with one thing. Test it. Learn from it. Iterate.

That's how I've approached every new technology across my businesses, and it's never failed me yet.

Ready to dive deeper into AI implementation for your business?

This is just the beginning of what's possible with AI in marketing. I test this stuff daily across multiple businesses, so you don't have to guess what works.

Want to see exactly how I'm implementing AI ads across Rocket Media and Digital Ignitor? Subscribe to the AI Brain newsletter, where I share weekly testing results, failures, and frameworks that actually work.

Remember: The best time to plant a tree was 20 years ago. The second-best time is now. The same goes for AI advertising preparation.