Stop Chasing Use Cases

Stop Chasing Use Cases

Two weeks ago, I wrote about the strategy/execution gap, especially use cases. Last week, I wrote about why most AI projects start in the wrong place — at the tool, instead of at Purpose. Today, I want to talk about the single most common form of “execution” that isn’t really execution at all.

Pilots.

I’m going to say something a little uncomfortable. Most AI pilots running inside companies right now are not experiments; they are procrastination with a project plan.

I know, because we’ve done it more than once. I’ll come back to that.

The pattern

A leader gets serious about AI. They do the strategy work. They commit budget. They approve three or four small pilots. One will focus on marketing, another on customer service, a third on finance, and maybe one on HR. The thinking is sensible. “Let’s try a few things, see what works, and double down on the winners.”

Six months later, the pilots are all still running. None of them has produced a result anyone can show to the board. None of them has a real owner. They all have “champions” who are doing this on top of their actual jobs. None of them has a budget for what comes after the pilot. They have not measured any of them against the same baseline twice.

This is pilot purgatory. This is what you’ve been hearing about. It’s the most costly type of AI fake progress in the industry.

Why “let’s try a few things” usually fails

It feels prudent. It is not prudent. Here’s what is actually happening when you spread one organization across five small pilots:

• Attention dilutes. Real change requires sustained focus from the leader and the team. Five pilots get one-fifth of the attention each. That’s not enough to push any of them through the inevitable hard middle.
• Ownership disappears. A pilot with five “stakeholders” and no single owner is a pilot that nobody is responsible for finishing. Everyone can claim a partial role; nobody has to deliver an outcome.
• The metric never lands. Each pilot has a different goal, on a different timeline, with different baselines, in a different part of the business. You can’t compare them. You can’t add them up. The board can’t act on them.
• There’s no organizational learning. Five tiny experiments at 20% effort produce five inconclusive results. One real experiment at 100% effort produces a result you can build on — even when the result is “this didn’t work, here’s why.”

The cure is the most counterintuitive thing I tell clients: pick one. Not the most ambitious one. Not the one with the loudest internal advocate. Not the one your CEO read about in a magazine. The one that has the clearest path to a real outcome.

The discipline of picking one

This is where TRIPS comes in. TRIPS is a scoring framework we built specifically for this decision and walk through in The AI-Ready Strategist. Think of AI as an outsourcing partner. TRIPS tells you what to actually put in the contract. Five criteria, one question each:

• T — Time. How much time is this task actually consuming? Add it up over a year. The “only 45 minutes a week” task is a full work-week a year. The tasks that feel fast — because they’re habitual — are usually where the real time is hiding.
• R — Repetition. Does the task follow the same pattern every time, or does every instance require unique judgment? AI loves patterns. This is the criterion that separates a real use case from a science experiment.
• I — Importance. This one works in reverse. The more important the task, the more human oversight it needs. A wrong word in an internal Slack summary is nothing. A wrong number in a board deck is career-defining. High-importance tasks can still use AI — but a human makes the final call.
• P — Pain. How much do people dread the task today? This is the emotional dimension most AI frameworks skip entirely. People don’t resist AI because they don’t understand the ROI. They resist it because they’re afraid it will take the parts of their job they love. Start with the parts they hate, and buy-in takes care of itself.
• S — Sufficient Data. Are there enough templates, past outputs, and documented examples for AI to learn from? If the knowledge lives in one person’s head, you’ve got a documentation gap to close before AI can help. Often, finding that gap is the first win.

When you put your candidate use cases through those five questions, the field collapses fast. Most of the things you were considering are interesting but not important. A few are important but premature. One or two are usually clear. They’re high-volume and repeatable. They’re also painful to do now and well-documented, so AI can learn from them.

That’s the one you ship.

The rest go on a list. They’re not “no.” They’re “not yet.” There’s a real difference between those two answers, and protecting that difference is part of the leader’s job.

The compound benefit of finishing one

Here’s the part I wish someone had told me earlier. When you finish one AI workflow — really finish it, all the way to “this is in production, the metric is moving, the team isn’t using the old way anymore” — three things happen that don’t happen with five half-finished pilots.

First, you have proof. You can walk into any meeting in your company and say, “this is what AI did for us in this workflow.” Skeptics get quiet. Budget gets easier. Recruiting the next champion gets easier.

Second, the team learns how to do this. Not the tool — the change. The hand-offs. The QA. The retraining. The data work. Those skills compound. The second workflow goes faster. The third goes faster still.

Third, you build the muscle of finishing. Most companies have plenty of starting muscle. Almost none have finishing muscle when it comes to AI.

Next week is the last issue in this series — the third of the three frameworks, and the one that makes or breaks everything else. The data work, and the 6C Framework we use to audit it, without which the best strategy and the best use case still produce garbage. Until then:

👉 The AI-Ready Strategist: trustinsights.ai/aistrategycourse

If you had to pick one use case, what would it be? Reply to this email or join the conversation in our Free Slack community, Analytics for Marketers!

– Katie Robbert, CEO

Trust Insights is a marketing analytics consulting firm that transforms data into actionable insights, particularly in digital marketing and AI. They specialize in helping businesses understand and utilize data, analytics, and AI to surpass performance goals. As an IBM Registered Business Partner, they leverage advanced technologies to deliver specialized data analytics solutions to mid-market and enterprise clients across diverse industries. Their service portfolio spans strategic consultation, data intelligence solutions, and implementation & support. Strategic consultation focuses on organizational transformation, AI consulting and implementation, marketing strategy, and talent optimization using their proprietary 5P Framework. Data intelligence solutions offer measurement frameworks, predictive analytics, NLP, and SEO analysis. Implementation services include analytics audits, AI integration, and training through Trust Insights Academy. Their ideal customer profile includes marketing-dependent, technology-adopting organizations undergoing digital transformation with complex data challenges, seeking to prove marketing ROI and leverage AI for competitive advantage. Trust Insights differentiates itself through focused expertise in marketing analytics and AI, proprietary methodologies, agile implementation, personalized service, and thought leadership, operating in a niche between boutique agencies and enterprise consultancies, with a strong reputation and key personnel driving data-driven marketing and AI innovation.