The short answer
AI content doesn't fail because AI wrote it. It fails because it says nothing only you could say. The fix is a system: build a master Cookbook about your business, keep one AI project that knows it inside and out, write about the real problems your customers bring you, and run every draft through a rinse-and-repeat round of questions until it sounds like you.
That's the process we use for our own content and for our clients. Here's every step.
Why bland content stopped working
For years, the play was simple. Write "What are the best general contractors in Greeley, Colorado?" Stuff in the keywords. Publish. Repeat.
That doesn't work anymore. AI tools can produce that article in ten seconds, so the internet is now full of it. And none of it helps anyone. People aren't asking AI who the best contractor is in the abstract. They're asking about the problem in front of them: why their basement floods every spring, whether their panel can handle an EV charger, what a fair price is for the job they just got quoted.
Google has been clear about where it stands. Its spam policies call out "using generative AI tools or other similar tools to generate many pages without adding value for users." The problem isn't the tool. It's content with nothing original in it.
Think about it logically. Five excavation companies in Dallas, Texas all want to rank for "excavation company Dallas." Each owner opens an AI tool and types, "Write me an article that helps my website rank for this term." That's all they give it. What comes back? Five versions of the same article.
When five competitors publish essentially the same content, why would Google or ChatGPT rank any of them? The chances are very low. The only thing that separates one company from the other four is the human touch: the stories, standards and answers only that owner has. That's why collaboration matters most on non-branded content, the searches where you're competing against everyone else's SEO slop.
Why AI search and Google move together
Google's AI Overviews are powered by Gemini, Google's own AI model, and they sit right on top of the search results. ChatGPT, Claude and Gemini are all trying to answer the same question: which business actually knows what it's talking about?
So the work cuts both ways. Content that earns a recommendation from AI search tends to be the same content that earns trust on Google. Build it right once, and it moves the needle in both places.
Step 1: Build your master Cookbook
Everything starts here. A Cookbook is a master document about your business, built by answering dozens of questions about how you actually operate: your competitive advantages, your value adds, what you say on the phone that books the job, the mistakes you see competitors make, what your pricing really covers.
Without it, every piece of content starts from a blank prompt, and a blank prompt produces the same generic answer it gives everyone else. With it, every piece starts from what only your business knows.
I won't pretend it's quick. Building a real Cookbook takes hours. Ours has taken somewhere between 15 and 20 hours, built up over time. Honestly, it would have been next to impossible without a dictation tool, which is Step 4.
"I don't have 15 hours." I get it, and that's a completely understandable reaction. But look at it as an opportunity cost, not a time cost. AI search is the Wild West, and it's going to be won by the trailblazers willing to put in the time. If you're one of the businesses that truly delivers value, with a unique approach to your service and how you talk to customers, your messaging is mission critical in this new landscape.
If 15 hours over a few months feels like too much, you have to come to terms with letting a competitor take that spot. Chances are, one of them will put in the time.
We laid out exactly how we build one, month by month, in The AI Search Engine Cookbook.
Step 2: Give your AI a permanent home
This is the step most people skip. Don't open a new chat every time you need content. Create one master project inside your AI tool, load your Cookbook into it, and save it as a tab on your desktop and your phone.
Then talk to it constantly. Every conversation teaches it more about your business. Over time, it stops sounding like a generic assistant and starts writing in your voice, almost like a messaging tool that knows your business as well as you do.
Step 3: Write about real problems, not keywords
Once your Cookbook is in place, content ideas come from your day, not a keyword list.
Whenever you or one of your best technicians is talking with a customer or a prospect about something they genuinely care about, and you have a perspective nobody else has, write it down. That's your next piece of content.
The homeowner who asked whether to repair or replace. The prospect who'd been burned by a contractor who didn't pull permits. The customer who wanted to know why your quote was higher than the other guy's. Each of those conversations is a page that answers a real problem, in your words.
Step 4: Talk it out
Most business owners aren't writers, and they shouldn't have to be. Plenty of the best tradespeople I know aren't comfortable at a keyboard at all. That's exactly why dictation matters.
Words per minute
The math is simple. The average person types about 35 words per minute, according to Ratatype's data from over 500,000 typing tests. When you're composing your own thoughts instead of copying text, research has measured it closer to 19. People dictate to a computer at about 105 words per minute (Human Factors International). That's the difference between a Cookbook that gets finished and one that never gets started.
Creating a Cookbook without dictation is, logically, very unlikely to succeed. Talk through each topic the way you'd explain it to a customer standing in their driveway.
I use Wispr Flow. It does a great job learning how I talk over time and cleaning it up into clear writing. Talk for five minutes, drop it into your master project, and ask your AI to turn it into a draft built from your Cookbook.
Step 5: Rinse and repeat
This is the step that turns a decent draft into content only you could have written. Once you have a draft you feel good about, send it to your AI and ask:
What five clarifying questions do you have that I can answer to make this article more valuable and feel more human and trustworthy?
Answer the questions, out loud if you want, and have it work your answers back into the piece. Then do it again. I usually run that loop once or twice more. Every round pulls out another real story, another specific number, another opinion only you hold.
This loop also takes the blinders off. It makes sure the content truly answers the questions your customer cares about, and it catches mistakes you can't see in your own writing.
Here's a real example. When we wrote our article "How Much Should You Spend on Google Ads? A Straight Answer," I dictated it, and it read great. Then the review rounds surfaced discrepancies in how we described our pricing and what results to expect. I didn't see them. The AI did, on the spot.
That's the point. If an AI tool can spot contradictions in seconds, assume search engines can too. Google's own guidance asks whether content "contains easily verified factual errors." Dictating without fact-checking is how credible businesses end up publishing content that undercuts their credibility.
That's the whole secret. The AI handles the speed, the structure and the fact-check. You supply the substance.
What this looks like in real businesses
We've used this process across the industries we serve, and it's consistently been the best way we've found to grow on AI search.

Hammerfall Roofing
Roofing, St. Charles, IL
6 to 8
qualified AI leads in one week, $0 in ads
Hammerfall Roofing is the clearest example. About four months in, Dean emailed us: 6 to 8 qualified leads in a single week, all from people who found him on AI platforms, with $0 in ad spend. By about six months, AI search had become his primary source of lead flow. Some clients get there sooner, but that's the kind of result that took a full Cookbook and consistent, human-led content to build.
Read the Hammerfall Roofing case study
Sander Law
Family Law, Glendale, AZ
Multiple
AI search inquiries every week
Sander Law went from zero AI search inquiries to multiple every week.
Read the Sander Law case study
Green Mountain Remodeling
Kitchen & Bath Remodeling, Lakewood, CO
5 to 10
qualified leads every month
Green Mountain Remodeling went from no organic leads to 5 to 10 qualified leads every month.
Read the Green Mountain Remodeling case studyNone of them got there with stock content.
The bottom line
AI is a tool in the hands of someone with real experience. Used that way, it produces content people want to read and AI search engines want to recommend. Used as a replacement for that experience, it produces slop.
- 1Build your Cookbook.
- 2Give your AI a permanent home.
- 3Write about real problems.
- 4Talk it out.
- 5Rinse and repeat.
Want to see the full six-month plan? Read The AI Search Engine Cookbook.
Frequently asked questions
Does Google penalize AI-generated content?
Not for being AI-generated. Google's spam policies target content mass-produced to manipulate rankings without adding value, however it's made. AI content built from real expertise and reviewed by a person is a different thing entirely.
What is a content Cookbook?
A master document about your business, built by answering questions about how you operate: your competitive advantages, your value adds, what you say to earn trust. Every piece of content starts from it, so it could only describe your business.
Do I have to be a good writer?
No. Talk it out with a dictation tool, let AI shape the draft, then answer its clarifying questions. Your knowledge is the hard part, and you already have it.
How long does it take to build a Cookbook?
Hours, built up over time. Ours has taken 15 to 20 hours so far. A dictation tool makes it realistic, since most people speak about three times faster than they type. If that still sounds like a lot, weigh it against the cost of a competitor getting there first.
How do I know what to write about?
Listen to your customers. Any time you or your team answer a question a customer genuinely cares about, and you have a perspective nobody else has, that's a piece of content.
Does content for AI search help on Google too?
In our experience, yes. Google's AI Overviews run on Gemini, and the qualities that earn AI recommendations, like original expertise, clear answers and trust, are the same ones Google rewards.
Want content that sounds like your business?
Let's talk about a marketing strategy that actually drives revenue.
Schedule a Discovery Call