How to Get Better Outputs From AI (The ROOTS Framework)
Tired of Generic AI? It’s Not the Robot, It’s Your Request
Sometimes, I get personally offended when people complain about AI. They’ll type something like, “Write me an Instagram caption for my new offer,” and then get upset when the output is a jumble of rocket ship emojis and exclamation points. It sounds like this: “Exciting news! I'm thrilled to announce my brand new offer that will transform your business! DM me to learn more.”
It’s generic, it’s soulless, and it sounds like everyone else on the internet, which means it sounds like no one at all. That frustration is real. You’ve probably tried it, gotten a garbage result, and decided that AI just can’t capture your voice or isn’t for you. You feel like you have to rewrite the whole thing, and at that point, what’s the point?
Here’s the thought we want you to hold onto: It wasn’t the AI that failed you. It was the prompt. You handed it a vague sentence and expected it to read your mind. The good news is that the prompt is the one thing you can control. The difference between getting robotic slop and getting content so good you barely have to edit it is the quality of your brief. Today, you’re going to get the exact framework we use to get outputs that sound like us every single time, so you can finally make AI feel like a partner, not a pain.
The Mindset Shift: AI Isn't a Vending Machine, It's an Employee
The biggest mistake people make is treating AI like a software, a tool, or a vending machine where you put in a simple command and get a perfect product out. It’s not. Think of AI as your newest team member. You would never expect a new hire to complete a task perfectly without any context, background, or training. You’d give them a brief, explain the goal, show them your brand voice, and set them up for success. The same rules apply here.
When you shift your mindset from operating a tool to briefing a team member, everything changes. You're not bad at AI. You’re just under-training it. To help you become a better manager to your new AI assistant, we developed the ROOTS framework. It’s a five-part system that ensures you give your AI everything it needs to give you back exactly what you want.
The ROOTS Framework: Your Guide to Better AI Prompts
This is the process we follow for every single request, and it’s going to make it so much easier to get what you need, faster.
R is for Role: Tell AI Who to Be
Before you tell AI what to do, you have to tell it who to be. This is the single most important move you can make to change the output. Without a role, AI defaults to a generic internet voice. With a role, it starts pulling its knowledge from a specific, expert lane. Think about it: if you ask a lawyer and a kindergarten teacher the same question, you’re going to get two very different answers. You need to tell it what hat it's wearing.
Don't say:“Write a blog post.”
Instead, say: “You are an SEO and AIO optimized blog post writer who writes in a plain, best-friend-to-best-friend language. You don’t just write blogs that are searchable; you write blogs that spark ‘aha’ moments and make readers want to come back for more.”
Or: “You are a conversion copywriter who specializes in writing for wellness practitioners.”
Get specific. You get to define your dream employee. The more detail you give about who you want the AI to be, the more it will step into that character and the less generic it will sound.
O is for Objective: Define the Goal
What are you actually trying to accomplish? So many people give AI a task, but not a goal. “Write a caption” is a task. “Write a caption that gets people to comment to boost this post’s reach” is an objective. This is a good practice for you as a business owner, too. It forces you to get clear on why you’re creating this piece of content in the first place.
Don't say: “Write a subject line.”
Instead, say: “Write a subject line that encourages the user to click through.”
Or: “Write an email to my business peer, Ashley. My goal is to get her curious about this idea and ultimately pitch her on a collaboration.”
When you don’t know what you want the output to achieve, AI is just guessing. Clarity on the human side of the prompt leads to a much more favorable response.
O is for Output: Specify the Format
If you don’t tell AI exactly what you want to get back, you’re going to get a 14-page essay when all you wanted was a list of ideas. AI, especially Claude, likes to be a little extra. You need to specify the format, constraints, structure, and even what to leave out.
Instead of just asking for ideas, specify: “Give me three subject lines under eight words each. Do not use any emojis or hashtags.”
Instead of a general request, structure it: “Write four short paragraphs. Do not use bullet points or em-dashes.”
You can even tell it what not to do: We’ve trained our AI to never start a podcast description with the phrase “Real talk” because it was overusing it.
This step saves you so much time in editing and formatting. It also saves you on usage credits because you’re not getting more content than you actually need.
T is for Tone: Hand It Your Voice
This is the step that turns generic AI slop into something that sounds human, and more specifically, like you. Your audience follows you for you. They recognize your language, your energy, and your way of communicating. If your content doesn't sound like you, it creates a disconnect. This is often the step people feel is impossible, but it’s actually the most trainable.
How to do it: If you don’t know how to describe your voice, don't guess! Find examples of writing you love, whether it's an email, a blog post, or even a social media caption you wrote, and paste it directly into the AI. Say, “Analyze this text and match this voice and tone.”
Pro-Tip: We took transcripts from our actual speaking voice on our podcast and fed them to our AI assistant. It constantly analyzes our speech patterns, cadence, and go-to phrases to learn our voice more deeply.
S is for Specifics: Give It the Context
The more context you can give your AI, the better the output will be. This is what takes your content from generic to deeply relevant. This includes specifics about your business, your offers, your audience’s exact pain points, the results you provide, and what makes you different.
When we were building the Life First Business Lab, we had trained our AI assistant on everything. It knew about the offer, the results it would get for people, the voice-of-client research we had done, and our personal stories and why this work is important to us. Because of that deep context, the AI helped us write sales pages, onboarding emails, and launch content that was 90% of the way there, only needing a few human tweaks.
You're Ready to Be a Great Manager
The next time you open an AI chat, I don't want you to see a blank box you have to fill with a magic phrase. I want you to see a new team member waiting for their brief. You're not trying to trick a machine; you're trying to communicate a goal clearly. When you take the time to walk through the ROOTS of your request—Role, Objective, Output, Tone, and Specifics—you set your AI up for success.
And I know what you’re thinking, “That sounds like a lot of time to write a prompt.” It is, at first. That’s why you save your prompts. Optimize them. The real time-saving happens when you have a prompt that works, and you can reuse it again and again.
If you want a cheat code to all of this, we invite you to check out the Life First Business Lab. Inside, you get a pre-trained AI assistant built on the ROOTS framework. After a quick onboarding to learn your specific business, it’s ready to help you write blogs, emails, and more, faster than you ever thought possible. You don’t have time to train a bunch of different assistants, which is why we created these to use in our own businesses. They are the employees we can’t live without.
