I’ve watched the same patterns play out over and over: business owner try to use AI for their marketing. They type in a prompt. The output comes back…it’s okay, but not quite right. So they try again. Different angle. Still not what they needed. They tweak the wording. Third attempt. Fourth. Fifth. By the time this roller coaster stops they’re frustrated. Exhausted. And quietly starting to believe something is wrong with them, saying:
“Maybe I’m just not good at this.”
“Maybe AI isn’t for people like me.”
“Maybe I’m asking the wrong questions.”
What I believe instead is this: nothing is wrong with the user. They’re just treating an intelligence system like it’s a tool. This distinction matters for anyone using AI in marketing, operations, research, or decision-making, especially as AI tools become embedded into everyday business workflows.
The problem with treating AI like a traditional tool
When you use a hammer, it does exactly what you direct it to do. Hit the nail. Drive it in. Done.
AI doesn’t work that way. AI doesn’t only execute, it interprets. It infers. It fills in gaps you didn’t know you left. When you ask it to “write a marketing email,” it’s making dozens of decisions you never specified:
- Who this is for
- What the goal is
- What tone is appropriate
- What matters most
- What should happen next
You didn’t answer those questions so the system answered them for you. That’s where outputs drift from intent. That’s where you get something that’s fine but not right. It’s not necessarily user error, it’s a faulty mental model. People are using an intelligence system like a hammer.
What it means to treat AI as an intelligence system (not software)
I come from a military intelligence background. And three things intelligence work teaches very clearly are:
- Information without context is dangerous.
- Ambiguity creates distortion.
- Systems must be led, not trusted blindly.
Those same principles apply directly to AI. AI doesn’t know what matters to you. It optimizes for patterns, not priorities. AI doesn’t understand your brand, your risk tolerance, or your audience unless you define them. AI doesn’t exercise judgment. It will confidently generate something that’s off-brand, misleading, or incomplete if you don’t set boundaries. The wrong question to ask is: “What can this tool do for me?”
The right question is:
“What am I asking this system to interpret, and how do I ensure it interprets correctly?”
That’s the difference between using AI and directing it.
What this looks like in practice
Lemme tell you about a real example from my work with a construction company client. The team wanted to use AI to support marketing and operational decisions. The request sounded simple: “Can AI help us figure out what to prioritize?” The issue wasn’t capability, it was direction. They were asking the system to generate recommendations without clearly defining:
- what success looked like
- who the decisions were for
- what constraints mattered (budget, capacity, timelines)
The AI did exactly what it was designed to do. It produced options. Lots of them. But instead of clarity the team felt overwhelmed. So we paused.
Instead of asking the AI to decide, we gave it a role: Analyze options based on stated priorities.
Surface tradeoffs. Flag risks. Stay within defined constraints. Same system, same data and a completely different outcome. The difference wasn’t the tool, it was the direction. This is what responsible AI use looks like in the real world.
Direction before execution
In my experience, people start with the output:
“Generate a marketing email.”
“Create social posts.”
“Build a content calendar.”
I believe that’s backwards. Consider starting with the direction….what outcome(s) are we trying to create? Who is the thing you’re creating for? What should this (the thing you’re creating) move toward?
Instead of:
“Write a donor thank-you email.”
Try:
“Write a 150-word thank-you email to a first-time donor who gave $50 to our youth literacy program. Tone: warm and personal, not corporate. Goal: increase the likelihood of a second gift.”
Stop expecting the AI tool to guess. Tell the tool what matters; lead from the front.
Signal over noise
AI can generate infinite content… does that mean it should? I don’t think so.
“Flooding the zone” with content is cool; the algorithms love it. But what if the goal isn’t more output for output’s sake? What if it’s clearer output that serves a specific purpose? When you treat AI like an intelligence system, you stop asking it to do everything and start asking it to do the right thing, well.
Instead of typing:
“Give me 10 social media posts.”
Try:
“Create 3 posts that address the top objections our audience has about working with a consultant.”
Fewer outputs and higher value. That tradeoff matters.
Human judgment is non-negotiable in AI systems
This is the line I draw with every client: You are the leader of the system. Don’t let the system lead you. AI is an assistant, not an authority. Human judgment is the control mechanism. That means:
- Editing outputs
- Verifying research
- Interrogating assumptions
- Deciding what gets used and what doesn’t
If an AI output was wrong, misleading, or off-brand… would you notice it before it gets published on your brand’s channels? If the answer is “maybe,” you’ve delegated too much. If not, the issue isn’t the tool. It’s the absence of direction, boundaries, and oversight. AI doesn’t replace thinking, it amplifies it. And intelligence systems MUST be led.
Why this matters now
I believe we’re past the “should we be using AI?” phase.
A lot of organizations already are without realizing how much judgment they’ve delegated in the process. The issue isn’t capability anymore. It’s control. Are you exposing sensitive data without realizing it? Diluting your brand voice? Letting a system optimize for speed instead of intention? The organizations that succeed with AI won’t be the fastest adopters. They’ll be the most intentional ones.
I work with individuals, businesses, nonprofits, and organizations to help them integrate AI responsibly, with clarity, guardrails, and strategy leading the technology. We’re not chasing trends, we’re practicing discipline.
If this distinction resonates, you already understand why this work matters.