The Why Behind This Rule
Look, my goal here isn't to try and ban you from using modern software, and it definitely isn't about protecting my job from a computer. I want you to use it.
This rule is about protecting the strategy we are investing our time and money to build .
Throughout my entire career, long before AI existed, I’ve watched projects get completely derailed by the exact same trap. It happens when a client takes a strategic layout or a copy draft and passes it around for outside feedback.
Note: Even your customers aren’t good to ask about your new website, they’re the ones that bought from your old website. If you really want constructive feedback about your marketing, ask the prospects who didn’t buy.
AI has completely supercharged this trap.
Because these chatbots have been heavily marketed as "the ultimate experts," it is easier than ever to type a quick prompt and get an instant critique.
AI often “sounds” smart.
Without even realizing it, people end up trusting a text predictor even more than their colleague or their spouse.
It’s the absolute quickest way to kill a strategy.
If you dilute our decisions and design, that we’ve made together, to satisfy a contextless algorithm, you lose all real conviction in the project.
When your business hits the inevitable sales or marketing bumps in the future, you won't have a solid foundation to stand on because the original strategy was quietly chipped away.
Let's protect the success of your new site, which means protecting the integrity of the choices we make together .
What Happens If It’s Not Followed
This trap is so incredibly easy to fall into, these guardrails are a firm boundary for how Dan operates .
Because these chatbots have been heavily marketed as "the ultimate experts," it is easier than ever to type a quick prompt and get an instant critique. This is what we will both agree to avoid doing.
This is NOT a ban on using AI, but an agreement to refrain from bluntly using it and trusting it as an equal partner with equal say.
If we find that project drafts, copy, wireframes, or designs are repeatedly being thrown into standard AI chats for critiques or final decision-making, we are going to pause the project and have an honest conversation to get back on track.
If we simply can't agree on keeping these strategic guardrails in place, that is completely fine, it just means we aren't the right fit to work together.
If it comes to that, we will cancel the project and part ways cleanly. There will not be any refunds for the strategy, consulting, or build hours completed up to that point .
Contract Terms
A clear and formal clause for this will be inside our actual contract.
We are working together to build a highly targeted asset that reflects your actual revenue goals. Let's make sure we don't let a generic chatbot strip that edge away.
How NOT to Use AI During Our Project
To protect our timeline, effort, and strategy that is all aimed at ensuring your site is built to convert real human buyers, we enforce a strict boundary on AI-generated feedback.
AI will Always Find a Flaw
The most dangerous thing you can do mid-project is drop a copy draft, wireframe layout, or design prototype into an AI and ask: "How can I make this better?" or "What is wrong with this?"
Here is the truth about how LLMs work: AI is programmed to be a cooperative assistant. If you ask it to find a flaw, it will always invent one to satisfy your prompt.
If you take a piece of text written entirely by an AI, throw it back into the same AI, and ask "How can I improve this?"—it will give you a list of critiques. If you apply those changes and ask again, it will give you additional critiques, often contradicting the previous round.
If you let an LLM run your review process, you will get sucked into an infinite loop of subjective tweaking, lose sight of your target audience, and stall your own project launch.
We design for objective business outcomes, not to satisfy an algorithm.
Do Not Use AI to Evaluate Layouts:
When you ask an AI to evaluate a design, it can only check if the layout matches the absolute most "common" patterns on the internet. It evaluates everything against a homogenized, generic baseline of general best practices.
But a chatbot doesn't sit in on our strategy sessions. It has no idea who your specific buyers are, what unique problems we've uncovered in discovery, or how we are intentionally layering sections to guide different stakeholders through your sales pipeline. Because it lacks that exact human context, it will look at a highly targeted strategy and try to pull it back toward a safe, generic middle.
Imagine we design a layout for a property management firm. We intentionally place a block of detailed text right below the header explaining your specific compliance certifications and response-time guarantees.
The Trap
You throw a screenshot of that mockup into an AI and ask if it’s a good layout.
The AI will instantly pull from its best web standards and tell you: "No, this section is too dense. Early sections should be minimalist.
You should replace this text block with three clean, 'modern' icons and bullet points for better conversions."
The Reality
The algorithm has zero context on your actual buyer.
It doesn’t know that your highest-value prospects are institutional property owners who have just been burned by a corner-cutting competitor.
Those specific buyers are actively scanning the page for proof that you take compliance seriously before they will even consider calling you.
If a layout ever feels off to you, or you’re worried a section might confuse a user, don't leave it to a standard AI window to guess. Let's look at it together during our next review meeting, check it against the specific targets we mapped out in discovery, and talk through how your actual, real-world buyers behave.
Do Not Use AI to answer Questions for you or make decisions:
It gives you an easy answer. It'll even sound incredibly smart. But, it completely skips the hard, necessary thinking that keeps your business unique.
Do Not Use AI to Dictate Your Decisions
The absolute worst question you can ask a standard LLM during this project is: "What should I do here?"
It is incredibly tempting to use AI as a shortcut for hard thinking, especially when we are working through structural business bottlenecks or refining your offers. But outsourcing your final judgment to a text predictor is an immediate mistake.
An LLM is a tool for compiling data and formatting ideas, it is not a partner on equal footing with human professionals. Here is why relying on it to make your project choices breaks down:
Example 1: The Commercial Contractor
Imagine we are defining the primary call to action (CTA) for a commercial builder's website. We are choosing between a generic "Get a Quote" button and a specific "Request an On-Site Blueprint Assessment."
The Trap
You ask a your favorite LLM, "Which CTA should I use?"
The AI will give you a "fantastic", split list of pros and cons, choose one at random, and write it in a way that sounds convincing.
The Reality
The machine doesn't know that your operations team is completely maxed out right now, that you want to filter out low-margin residential requests, or that an "On-Site Assessment" aligns perfectly with how you historically land your biggest six-figure commercial projects.
Example 2: The Multi-Persona B2B SaaS Page
Imagine we've built a homepage draft for a long-life sales cylce B2B software platform. Through our strategy sessions, we identified that closing a deal requires converting three entirely different personas who read the page with completely different motives:
The Trap
You paste the page copy into a blank AI chat and ask: "Is this page layout optimized, or should I change the order of the sections?"
The standard LLM will immediately default to generic best practice advice.
"This page is too long and disjointed. You should tighten the copy, remove the technical steps, and keep the text completely unified around a single persona to maximize engagement."
The Reality
If you follow the AI's "best practice" advice, you break the intentional, multi-tiered structure we built.
By stripping out the technical details, the end worker leaves the page because they don't see how the software actually fixes their problem.
By stripping out the team-focused language and social proof, the C-Level blocks the purchase because your company looks like a low-impact commodity instead of a serious partner they can trust.
The AI recommends best practices and won't take into consideration (unless you meticulously provide it) the messy reality of how three different human beings buy software together.
Use AI to help you clear up a messy brain dump, layout comparison charts, or organize raw information. But do not let it do your thinking for you, and do not let it make the final call.
We are building a custom asset to reflect your actual operations. When a tough strategic decision comes up, let’s bring it to our live refinement rounds and solve it using our combined human context.
The Copywriting Trap: The Context Problem
When it comes to reviewing or changing the words on your site, there is a right way and a wrong way to use an LLM.
Most business owners accidentally choose the path that breaks their messaging simply because no one has shown them what happens under the hood.
The Path to Avoid (Pasting into a Blank Chat):
This happens when you open a standard, blank AI window, paste in a draft of the website copy we delivered, and type something like: "Make this sound better" or "Make this more engaging."
Because that blank chat window has absolutely no context about your business goals, your target customer's frustrations, or your specific market, it can only pull from generic best practice patterns. It will systematically strip out the human edge, replace it with sterile marketing fluff, and make your business look like a commodity or miss the opportunties we're planning to target.
How We Actually Use It (Heavy Context + Constant Guardrails):
When Dan uses AI to help draft or refine copy, it's not treated like a magic button. It's treated the same way Dan has approached strategic copy for years, just faster.
Before a single prompt is written, a custom notebook/project (using tools like NotebookLM or specialized AI projects) is created. It's fed everything we've gathered: your raw background, our strategy call notes, every piece of internal data we could uncover, and Dan's own library of design and marketing frameworks purposely built for AI research and assistance.
Significant effort is dedicated to heavily iterating when using LLMs, explicitly pushing back when it slips into generic marketing speak, correcting its tone, and forcing it back into focus.
Most importantly, every single line is evaluated, iterated, and reworked against real human experience. Responses are never simply taken copy/pasted onto a page.
If you take an engineered sales narrative that we spent weeks blueprinting and run it through a standard, contextless AI prompt, the conversion strength of that copy will be completely diluted.
If a section of copy doesn't feel right to you, don't worry about trying to rewrite it with a prompt. Tell me what business objective we missed or what needs to change, and we will refine it together.
How YOU Can Use AI Productively
We don't want you to turn off your tools. If you want to use AI to support your role in this project, here is where it shines:
Then, send that raw material to us .
Our Stance
When you partner with Dan, together, our goal is to deliver a better-looking, high-converting touchpoint that reflects the true scale of your business.
Let's use technology to speed up data collection, but let's leave the strategic decisions to the humans in the room.
We guide the project based on your real-world goals, the pain you’re needing to resolve, sales psychology for your specific customer base, and years of design experience… not prompt engineering .