Why Problems Make the Best Use Cases
“Let’s have a brainstorming session to come up with some good ideas for AI.” Or maybe, “What’s a good AI use case for us?”
These are the questions most leaders start with, and they sound logical. But they often lead to generic answers like chatbots, predictive analytics, or automation, or they lead to projects with large, ill-defined scope that explain why 95% of AI pilot projects are failing. Those are categories, not real opportunities.
The truth is, AI creates the most value when it’s applied to understanding and solving specific problems. Problems define the pain point, the urgency, and the measure of success. Without a well-defined, lucid problem to solve, an “AI use case” is just a science project.
Problems Give AI Purpose
For AI to create real value, a problem has to be more than a vague complaint. It needs enough detail to explain what’s happening, who’s affected, and what success would look like. The sharper the problem, the clearer the path for AI.
Take a customer problem. Instead of saying, “support is bad,” a sharper definition would be: “Our customer support team in the U.S. market takes three days on average to respond, relying on email logs and ticketing data, but compliance rules require us to resolve certain issues within 48 hours.” This points directly to AI-powered intake and triage.
Or consider an operational problem. Not just “data entry is annoying,” but “our sales team spends 30% of their time retyping order details between ERP and CRM systems, using contract and transaction data, under strict accuracy standards that cause delays when errors occur.” That shows exactly where process automation can help.
When you begin by framing problems this way, the value of solving them is clear and measurable. That’s why Kendall’s principle says: problems fuel AI.
Dig Deeper: Learn how Kendall's AI Jumpstart supercharges your AI efforts in just 30 days
Context Turns Problems Into AI Fuel
Once you’ve identified a problem, the next step is to give AI the context it needs to solve it. Problems aren’t just headlines like “reporting takes too long” or “support tickets pile up.” To be useful, they need to be unpacked into their full picture.
That means clearly defining the components of the problem:
- Who is impacted - team members, customers, or partners.
- Where it happens - locations, workflows, or specific markets.
- What data is involved - inputs, sources, and formats.
- What constraints apply - compliance rules, policies, budgets, or timelines.
At Kendall, this structured approach leads to something we call a context bill of materials. By structuring problems this way, you transform vague complaints into precise, AI-ready inputs.
When you aggregate and deliver this high-quality context to AI, you’re not handing it a generic challenge. You’re giving it the full explanation of the problem and the environment it lives in. Imagine asking a new teammate, no matter how smart, to solve a persistent enterprise problem without explaining all of the specifics of that problem to them in detail. Assembling, organizing and training your AI on this problem-focused context is what ensures it doesn’t just generate ideas but actually succeeds in solving the problem, consistently, repeatedly and for the long-term.
Why Brainstorming Use Cases Fails
Brainstorm sessions often chase shiny objects. Someone suggests “let’s build an AI dashboard” or “let’s try generative design,” but without a grounding problem and its context, it is difficult to measure success.
By contrast, when you start with a problem and build its context bill of materials, success is easy to define: Did AI reduce costs? Did it shorten cycle times? Did it improve customer outcomes in the specific place the problem lived? That level of clarity is what separates organizations that dabble in AI from those that scale it.
The Takeaway
If you’re stuck trying to “find use cases,” flip the script. Don’t chase ideas. Collect problems, break them down into their full context, and then deliver that context to AI.
Problems fuel AI, but context makes them solvable. Together, they give AI direction, urgency, and measurable impact. That’s not generic experimentation, it’s focused execution, and it’s where the real business value lives.