Stop building those AI Agents!
Start Solving REAL Problems.
Lately, I’ve been in a lot of rooms where the first question someone asks is:
“Are you building agents yet?”
OR
“What AI agents have you built?”
And honestly? That’s when I know we are already off track.
When we lead with buzzwords - “agents,” “GPT,” “LLMs,” and “autonomous systems,” we are not starting with the problem.
We are starting with the hype.
And hype might sound exciting in a strategy deck.
But it rarely delivers anything useful to our team.
Plus, it often creates a false sense of urgency because it leaves you with FOMO.
Like - OMG - I am late to the party!
The Real Work? It's Not Trendy.
Everyone wants the shiny stuff. But here’s what we should be asking instead:
What’s broken?
Who’s frustrated?
Where are we wasting time or losing trust?
What do people wish they could do, but can’t?
If we don’t start there, we risk solving nothing - just spending budget on tools that sound powerful but feel pointless.
Because here's the truth:
AI is NOT the solution for all our problems.
A Framework That Actually Works
Here’s what I tell every HR team I work with:
Talk to people.
Understand the problem.
Show them how some AI tools work and explain the fundamentals of AI.
Then, and only then, decide if AI is the right fit.
Sometimes it will be.
But just as often?
The best fix is something far less glamorous:
Removing clutter and waste from a process, such as unnecessary approvals that waste time and add no value.
Asking questions like: Why are we doing this process the same way we did 10 years ago?
Better naming conventions for files, reports, folders, etc.
Cleaner documentation that is accessible to all stakeholders.
A decision tree taped to a wall
A Kanban board.
A to-do list
If you jump to “AI” before you have clarified what’s actually needed, you are skipping the most human part of the process: understanding each other.
Know the Tech. Lead with the Problem.
Now, to be clear, I’m not saying “don’t learn about agents.”
It’s great to understand what is possible.
We should know what agents can do.
We should explore where LLMs are heading.
We should be curious about RPA - robotic process automation.
Just don’t lead with them.
Keep them in your back pocket - not your opening line.
When a real problem shows up that matches the strengths of that tech?
Pull it out. Use it. Build something brilliant.
But don’t reverse-engineer problems just to justify the trend.
Not Every Agent Is Agentic
Let’s talk a bit more about “agents.”
There’s a lot of heat around the idea that autonomous agents will run our processes, make decisions, and take work off our plates.
And yes, it’s cool tech.
Just because it’s labeled “AI-powered” doesn’t mean it can solve YOUR problems.
Just because it’s called an “Agent” doesn’t mean it works for you and solves YOUR problems.
Real value doesn’t come from technical labels.
It comes from real-world outcomes.
Sometimes the Best Solution is Boring
And you know what?
That’s okay.
Sometimes the most effective solution isn’t innovative - it’s just a simple, streamlined process.
Clarity doesn’t need a chatbot. Trust doesn’t come from a fancy agent. Sometimes the solution isn’t artificial. It’s just honest.
The goal isn’t to “use AI.”
The goal is to reduce friction.
To free up time.
To make better decisions.
To improve the lived experience of work.
That’s it.
A Simple Gut Check for Every AI Project
Before you build, buy, or brainstorm your next AI idea, ask these three questions:
What’s the actual problem we’re trying to solve?
Who’s experiencing it?
Is it a skills gap? A process bottleneck? Would training or better communication solve it faster?If AI didn’t exist, how would we try to fix this?
If the solution still points to AI after all that?
Go for it. You are probably on the right track.
But if it doesn’t - congrats. You just avoided a lot of wasted time, money, and “cool features” no one asked for.
Don’t hand someone a power tool when all they need is a pen.
Final Thought: Start with People, Not Tech
Some of the best AI use cases I’ve seen?
They don’t even mention AI.
They start like this:
“This part of my job is painful. I spend 4 days doing this every single month. Can we fix it?”
That’s the energy we need more of.
The future of work will be shaped by the people who asked better questions before they picked a tool.
Rethink!