AI Is A Tool, Not the Objective

3 Real-World Industry Use Case Examples

1. The Ship Builder

A little over a year ago, I lost an opportunity to help a shipbuilder shorten the time it takes to deliver ships to the military. The potential ROI was substantial because the core problem, job shop scheduling, is a classic hard problem where mathematical optimization works extremely well.

The reason we lost? The budget was pulled because “Leadership wants us to focus on cutting edge AI – something that uses LLMs or maybe drones. Optimization isn’t seen as new enough technology.”

I continue to see a lot of that thinking. Statements like, “IT made this big investment in some massive hardware and now we need a way to consume it.”

Last I checked, a company’s goal (and its fiduciary responsibility to shareholders) is to make money, save money, and reduce risk. AI, drones, robots, and other modern flashy tech are absolutely opening up new ways to do that. But too often, the tail (the technology) ends up wagging the dog (the business objective).

2. AI-Powered CRM

One place I personally see a lot of AI hype is sales forecasting and pipeline management. Many B2B companies have invested in AI & ML-driven forecasting tools that ingest CRM data, emails, and activity signals to predict deal close probabilities. In practice, these models often underperform a simple Excel forecast built on historical win rates by segment, with manual adjustments for known deal-specific factors like budget approval, legal review, or executive sponsorship.

The AI struggles because CRM data is incomplete, sales behavior is inconsistent, and deal dynamics change quarter to quarter. A weighted spreadsheet with explicit assumptions is often more accurate, easier to audit, and far more trusted by sales leadership.

Starting with the most sophisticated technique isn’t a sign of technical maturity. It’s often a sign that the problem hasn’t been clearly defined. At the same time, “Start simple” is not the same thing as “stay simple.”

“Start simple” is not the same thing as “stay simple.”

3. The Greeting Card Company

Earlier in my career, I worked with a large greeting card retailer that issued an RFP for a business rules management system (BRMS) to optimize how cards were laid out on store shelves. They had rules like: high-margin cards should sit at eye level; humorous cards shouldn’t be placed next to sympathy cards, etc.

At the same time, they were solving a much harder problem: assortment optimization. Out of roughly 10,000 cards in their corporate inventory, which subset should go to each store based on local demographics and historical sales? They were already handling this with heuristics and asked us, as the BRMS vendor, to ignore that part of the problem and stay in our (business rules) lane to address only the layout problem.

We pushed back. Our view was that rules alone would cap the upside. By combining a rule engine with an optimization solver, we could jointly decide which cards belonged in each store and where they should be placed, maximizing expected return per store rather than merely enforcing placement logic.

The implementation was more complex and took longer and cost a bit more than anticipated, however, two years later, their CIO thanked us for supporting them in delivering the most successful IT project of the year.

The lesson isn’t that more or less complex systems are always better. It’s that complexity earns its place when simpler approaches leave meaningful value on the table and when the business objective is clear enough to justify the tradeoffs.

---

About the Author

 

Nevra Ledwon is a Managing Director at SimpleRose with more than 25 years of experience helping innovation-minded organizations apply advanced technology to achieve measurable business impact. Her work has long focused on data, analytics, optimization, natural language processing, computer vision, and AI–well before these capabilities were widely recognized or labeled as such.

She supports SimpleRose clients in solving complex planning and decision-making challenges, and is always open to a conversation if you have a problem or project you’d like to discuss.