AI is getting more capable. But better AI doesn’t automatically produce better work. The quality of what you get still depends on how well you communicate what you actually need.
In this Leveraged Marketing Mastermind session, Patrick Ferry breaks down a practical framework for getting more useful results from AI—and turning what you already know into systems that can be executed more consistently.
WATCH THE SESSION
[EMBED LMM REPLAY VIDEO HERE]
Turn the Lesson Into Action
Watching the session is one thing. Building the framework into the way you actually work is another.
Download the CRIT AI Playbook to walk through the framework step by step and apply it to a real process in your business.
DOWNLOAD THE CRIT AI PLAYBOOK
Context. Role. Interview. Task.
Use the playbook to identify what AI needs to know, how you want it to think, what it should clarify before starting, and what a successful outcome actually looks like.
Better AI Starts With Better Instructions
There’s a lot of attention right now on which AI model is smartest, which new tool just launched, and what AI might be able to do next.
Those things matter.
But there’s a more immediate question:
Are you actually communicating with AI in a way that gives it a chance to do great work?
You can have access to an incredibly capable model and still get a mediocre result.
Why?
Because AI can’t automatically know everything that is in your head.
It doesn’t necessarily know your standards. It doesn’t know what you’ve already tried. It doesn’t know what a successful outcome looks like to you. And it may not know which details matter most unless you tell it.
That is where CRIT comes in.
The CRIT Framework
CRIT stands for:
C — Context
R — Role
I — Interview
T — Task
It is a simple framework for creating a better working relationship with AI.
Instead of throwing a short prompt into ChatGPT and hoping the answer is useful, you deliberately give the AI what it needs to understand the situation, think about it appropriately, clarify what is missing, and then execute a clearly defined task.
Let’s break it down.
C — Context: Give AI the Information It Needs
Most people dramatically under-explain what they want.
They type a few sentences and expect AI to somehow understand the business, the client, the situation, their standards, their experience, and the outcome they have in mind.
Then they’re disappointed when the answer feels generic.
The first step is to provide context.
That might include your documents, SOPs, checklists, examples, transcripts, videos, project information, previous work, or other relevant resources.
But context isn’t limited to files.
It also includes what you know.
Your experience.
Your opinions.
Your concerns.
What you’ve already tried.
What you’ve learned.
What you believe a great outcome should look like.
One practical shift is simple:
Stop typing so little. Start explaining more.
Use voice when it makes sense. Talk through the problem. Give AI the background you would give a capable person if you were bringing them into the project.
The better it understands the world surrounding the task, the better chance it has of helping you.
R — Role: Tell AI How You Want It to Think
Once AI understands the situation, give it a role.
Ask yourself:
Who would I want helping me solve this problem?
For a marketing campaign, perhaps you want an expert copywriter.
For a complicated implementation, perhaps you need an expert project manager.
For a business decision, you might want it thinking like a CEO or chief of staff.
For content, you might need a social media strategist.
For search visibility, you might want an SEO or AEO perspective.
The role changes the lens through which AI evaluates the context you’ve provided.
Instead of simply asking:
“What should I do?”
you are giving it a clearer perspective:
“Evaluate this as an expert project manager.”
or:
“Approach this as an expert copywriter.”
When appropriate, you can also point it toward a specific framework or expert perspective you want applied.
The goal isn’t to make the prompt sound sophisticated.
The goal is to give AI a useful point of view from which to think.
I — Interview: Make AI Ask Before It Acts
This may be the most overlooked part of the framework.
Once you’ve provided context and assigned a role, don’t immediately ask for the final answer.
Give AI an opportunity to identify what is missing.
But there is an important distinction.
Simply saying:
“Ask me any questions you have.”
can lead to an unnecessarily long interrogation.
Instead, constrain it.
For example:
Ask me the three most important questions that would determine the success of this outcome.
Now you’re forcing AI to prioritize.
What does it genuinely need to know?
What assumption could change the result?
What information is missing that would materially improve the work?
Answer those questions before moving forward.
This creates a checkpoint between what you think you’ve communicated and what AI actually understands.
T — Task: Define What “Done” Looks Like
Now tell AI exactly what you want it to do.
This sounds obvious, but it’s one of the easiest places to lose control of the interaction.
You can provide tremendous context, assign the right role, answer every question—and then finish with:
“Okay. Go.”
And suddenly you have pages of information you never wanted.
Instead, define the outcome.
Do you want a recommendation?
A document?
A reusable process?
A project plan?
A summary?
An action performed?
A breakdown of responsibilities?
Be specific.
For example:
Tell me what AI can handle, what my staff should handle, and what only I should handle.
That’s dramatically clearer than simply asking AI to “help.”
The clearer your definition of done, the more useful the output becomes.
The Bigger Opportunity Isn’t Prompting. It’s Systems.
CRIT becomes even more interesting when you stop thinking about AI as something that simply answers questions.
The larger opportunity is to look at the processes you already understand and ask:
Could this become a system?
You may already have checklists that consistently produce good outcomes.
You may have processes you’ve refined through years of experience.
You may know exactly how a task should be performed, even if it still requires a lot of manual work.
That knowledge is valuable.
As AI tools become more capable and increasingly able to work with other tools and information sources, clearly documented standards and processes become even more important.
Instead of repeatedly asking AI to solve isolated problems, you can begin asking:
How do I teach the system what good looks like?
Don’t Automate Something You Don’t Understand
There is a temptation with every new AI development to immediately search for another tool.
Another subscription.
Another automation.
Another integration.
But before adding more technology, look at what you already know how to do well.
Pick a process where you understand the desired outcome.
Document the steps.
Define the standards.
Then use AI to help you determine what portions could potentially be assisted, accelerated, or systematized.
The goal isn’t to automate everything overnight.
The goal is to start with something you understand well enough to evaluate whether the result is actually good.
AI Can Also Make You a Better Leader
There is another lesson hidden inside this framework.
Working effectively with AI is an exercise in delegation.
If AI gives you something completely different from what you expected, it’s worth asking:
Did I actually explain what I wanted?
That same question matters when working with people.
Did you provide enough context?
Did you make the person’s role clear?
Did you give them an opportunity to ask important questions?
Did you define what a successful finished product looks like?
CRIT isn’t only an AI framework.
It can reveal where your own communication needs to become clearer.
And because you can practice these interactions repeatedly with AI, you get an environment where you can improve those delegation skills quickly.
Start With One Process
You don’t need to rebuild your entire real estate business around AI.
Start smaller.
Choose one thing you already know how to do well.
Maybe it’s a marketing campaign.
A listing process.
A client follow-up workflow.
A content process.
A recurring administrative task.
A checklist you’ve used dozens of times.
Then work through four questions:
CONTEXT — What does AI need to know?
ROLE — Who do I need AI to be?
INTERVIEW — What does AI still need to learn from me?
TASK — What exactly do I want accomplished?
Once you’ve done that, ask the bigger question:
What can AI do, what should my team do, and what can only I do?
That’s where this gets interesting.
Because the advantage isn’t simply knowing how to use the newest AI tool.
The advantage is taking what you already know, defining what good looks like, and turning that knowledge into systems that can be executed consistently.
Put CRIT to Work in Your Business
The CRIT AI Playbook gives you a practical worksheet for applying the framework to one real process in your business.
Don’t just learn the framework.
Use it.
