Stop Asking AI to Do the Work: Teach It How to Work With You

There is no shortage of advice telling businesses to use AI. Write your social media posts with it. Draft your emails. Create your blog articles. Build your business and marketing plans. Develop a presentation. Brainstorm your next campaign. The list goes on.

And yes, AI can do all of those things. It can be an incredibly powerful tool for a brokerage, a proptech company, a real estate vendor or almost any other business operating in and around the industry.

But there is a problem I am seeing more and more often. Businesses are asking AI to create content without giving it enough information to create their content.

The result is exactly what you would expect. Generic articles, predictable social posts, the same tired phrases and content that could have been produced for almost any company in the industry. Worse, because so many businesses are now using AI in the same way, we are starting to see a lot of content that sounds remarkably similar.

That is not an argument against using AI.

It is an argument for using it better.

After more than 20 years working in marketing, communications and business strategy, I have spent a great deal of time thinking about how AI can support the work we do. The biggest lesson for me is that AI works best when you stop treating it like a machine you give instructions to and start treating it more like a member of your team. And just like any new member of your team, it needs to be properly onboarded.

AI does not know your business like you do

One of the biggest mistakes people make with AI is assuming that because it has access to an enormous amount of information, it somehow understands their business. It doesn't.

AI may know a great deal about real estate, marketing, technology or your particular industry, but it does not automatically know your brokerage. It does not know why you made certain business decisions, what your clients value most, what your brand sounds like, what makes your company different or which industry clichés make you cringe.

If you type, "Write a LinkedIn post about our brokerage," you have essentially asked AI to fill in all of those blanks itself. So it will.

And, because it does not know what makes your business distinctive, it will fall back on the language and patterns it has seen countless times before. You end up with something technically acceptable, but completely forgettable. This is why I believe the quality of your AI output is often determined before AI writes a single word. The better the information and direction you give it, the better chance you have of getting something that actually sounds like you.

Think of a prompt as a creative brief

Anyone who has worked with a good marketing team knows that a creative brief matters.

If you tell a designer, "Make me an ad," you are probably going to have a frustrating conversation. What are we advertising? Who is it for? What are we trying to achieve? What does the brand look like? What is the offer? Where will the ad appear?

The same principle applies to AI. Instead of simply telling it what you want produced, give it the information it needs to make good decisions. For example, imagine a brokerage wants to promote a new agent education program. A basic prompt might say:

"Write a social media post promoting our new agent education program."

There is nothing technically wrong with that prompt. There is just very little there for AI to work with. A much stronger prompt might explain that the brokerage is an independent Ontario brokerage with 100 REALTORS®, that the program is designed for newer agents, that the brokerage's differentiator is hands-on mentorship, that the tone should be knowledgeable but approachable, that the goal is to encourage agents to register, and that the post should sound like it came from the broker rather than a marketing agency.

Now AI has context.

I generally think about a strong prompt through a handful of basic questions:

Who should AI be? Give it a role or perspective.

What does it need to know? Provide the relevant business context.

Who are we talking to? Be specific about the audience.

What are we trying to accomplish? Explain the objective, not just the deliverable.

What should it sound like? Describe the tone and provide examples where possible.

What should it avoid? Be clear about language, claims, assumptions or approaches that are off limits.

What do we want back? Define the format, length and other requirements.

You do not need to include all of these elements every time. A simple task does not require a three-page prompt. But when you are asking AI to help with something important to your business, context matters.

The most important step may happen before the prompt

If you want AI to consistently produce useful content for your business, don't start by asking it to write. Start by teaching it. This is particularly important if you are going to use AI regularly for marketing and communications.

Give it your brand guidelines. Share your website copy. Provide examples of your previous content. Explain your audience. Tell it about your products or services. Explain your positioning and your values. Most importantly, show it what your business actually sounds like. You can even ask AI to analyze your existing material before it creates anything new.

For example, you could provide several pieces of content that you feel represent your brand particularly well and ask:

"Analyze these examples and identify the characteristics of our writing style. Look at tone, sentence structure, vocabulary, level of formality, use of humour, calls to action and the types of language we tend to avoid. Create a style guide that you can use when creating future content for our business."

That is a very different approach from simply saying, "Write me a blog post."

You are beginning to build a working relationship with the tool.

You are teaching it how your business communicates.

And, that is where the real opportunity starts.

Don't just tell AI what you like. Tell it what you don't like.

One of the most useful things you can do when training AI is provide negative examples.

Most businesses are very good at saying what they want. They are less likely to articulate what they don't want. Yet those boundaries can be incredibly useful.

Better still, show it examples.

You might say:

"These three examples sound like us. These two do not. Analyze the differences and explain what you should do differently going forward."

That kind of feedback is far more useful than repeatedly asking AI to "make it sound more human." Human is not a writing style. Your business has a writing style. The more clearly you can define it, the more likely AI is to reproduce it.

Your expertise is the ingredient AI cannot manufacture

This is particularly important for real estate businesses. There is a tremendous amount of generic information about real estate available online. AI has absorbed patterns from an enormous amount of that information, which means it can produce a perfectly reasonable article about buying a home, recruiting agents, social media marketing or the benefits of technology in seconds.

But reasonable is not the same as distinctive. What makes a piece of content valuable is often the perspective behind it. For a broker, that might be 15 years of experience dealing with agents, consumers and the realities of running a brokerage.

For a PropTech company, it might be the specific problem the founders encountered in the industry that led them to build their product.

For a real estate law firm, it might be the recurring issues they see clients misunderstand.

For a marketing company, it might be a lesson learned from managing hundreds of campaigns.

That expertise needs to be brought into the AI conversation. Instead of asking AI to write an article about "five ways brokerages can improve agent retention," give it the five things you have actually learned about agent retention. Tell it what you have seen work. Tell it what has failed. Give it examples. Explain your point of view.

Then ask AI to help you organize and communicate that expertise.

That is where AI becomes an amplifier rather than a replacement for your thinking.

Ask AI questions before asking it for answers

One of my favourite ways to work with AI is to make it slow down before it starts producing.

If I am working on a strategy, I don't necessarily want AI to immediately give me a strategy.

I might first ask it to identify what information is missing.

For example:

"Before developing this marketing strategy, ask me the 10 questions you need answered to make your recommendations more specific to our business."

That changes the conversation. Instead of AI guessing, you are giving it an opportunity to identify gaps in the brief.

You can do the same thing with content. If you want to write an article about a challenge facing your industry, ask AI what questions you should answer to make the article genuinely useful. Then provide your answers. You may find that the resulting article contains much more of your actual expertise and perspective. This is one of the reasons I prefer thinking of AI as a collaborator rather than a content vending machine. The best results often come through a conversation.

Don't let AI become the voice of your company

There is a temptation, particularly for smaller businesses and startups, to let AI take over content creation because it is fast. I understand why. Marketing takes time. Founders are busy. Brokerage owners have a hundred other things competing for their attention. A tool that can produce a blog post in seconds can feel like a solution.

But speed should not be confused with strategy.

If every company in your sector is using the same AI tool, asking it the same generic questions and publishing the first answer it produces, you are not creating a competitive advantage. You are creating more noise.

Your audience may not know that AI wrote your article, but they will recognize content that feels interchangeable. The irony is that AI can actually make your content more distinctive, but only if you put your own knowledge into the process.

Use AI to organize your thoughts. Challenge your assumptions. Identify gaps. Suggest alternatives. Repurpose a strong piece of content into different formats. Help you research a topic. Create a first draft that you can improve. But don't hand over the entire responsibility for your company's voice to a machine that knows nothing about your business unless you take the time to teach it.

And, remember that AI can be wrong

There is another reason not to simply copy and paste what AI gives you.

It can be wrong. AI can produce information that sounds incredibly confident even when the underlying information is incomplete or incorrect. That matters enormously in real estate, where regulations, contracts, market statistics, financial information and other details can have real consequences.

AI should not be your final authority on matters that require professional, legal, regulatory or factual verification. If you are giving AI source material, tell it to work from those sources. Ask it to identify where information is missing. Ask it to flag assumptions. And always have the appropriate person review information before it is published or relied upon.

The goal is not to remove human oversight.

It is to make human oversight more efficient.

The goal isn't more AI content. It's better business.

I think this is the part that often gets lost in the excitement around AI.

The goal should not be to produce 30 social media posts instead of 10.

It should not be to publish a blog every week simply because AI makes it easy.

And it certainly should not be to replace every piece of human thinking with an AI-generated answer.

The real opportunity is to use AI to help your people spend more time on the work that requires judgement, creativity, relationships and expertise. For a brokerage, that might mean giving a marketing or operations team more capacity. For a startup, it might mean helping a small team compete with organizations that have significantly more resources. For a real estate vendor, it might mean turning years of industry knowledge into better educational resources, communications and customer support.

AI can help you move faster.

But moving faster in the wrong direction is not particularly helpful.

Start with one thing

You don't need to create an elaborate AI strategy tomorrow.

Choose one area of your business where you are already spending a significant amount of time. Maybe it is social media. Maybe it is internal communications. Maybe it is creating training materials, responding to common questions or developing marketing campaigns.

Then take the time to teach AI about your business before asking it to produce something important.

Give it context. Give it examples. Explain your audience. Define your voice. Establish boundaries. Tell it what good looks like. Then start prompting.

Review what it produces. Tell it what is right and what is wrong. Refine the instructions. Add information. Try again. Over time, you will build something much more valuable than a collection of clever prompts. You will build an AI workflow that actually understands how you work.

And that is the difference between using AI to create content and using AI to strengthen your business.

The technology is available to almost everyone.

Your experience, your perspective and your knowledge of your customers are not.

Need help figuring out where AI fits into your business?

At The Real Estate Source, we work with brokerages and businesses that support the real estate industry to identify practical opportunities for AI across marketing, communications, operations and education.

That can include helping your team develop better prompts, establish AI workflows, train an AI tool around your business and, most importantly, understand where human expertise still needs to lead the process.

If you are using AI but finding the results generic, or you know there is more potential but aren't sure where to start, feel free to reach out and book a complimentary consultation.