-->

Register Now to SAVE BIG & Join Us for KMWorld 2025, November 17-20, in Washington, DC.

What Problem Is AI Actually Trying to Solve?

Article Featured Image

Case Study: The Oil and Gas Company That Fixed the Wrong Problem

A telling example comes from a major oil and gas firm that wanted to replace its legacy KM system. Employees hated it, citing slow access to large files in remote locations. Consultants (including the author of this article) identified the real issue: poor bandwidth at local sites, not the KM system itself.

Despite warnings, the company spent millions on a new system—only to end up running both the old and new systems in parallel, with the same remote access issues persisting. The actual fix—upgrading from copper wire to satellite connections—had been proposed by remote workers for years but was ignored.

This illustrates a common pattern: Organizations use technology to treat symptoms rather than root causes. AI, if misapplied, risks becoming the latest iteration of this mistake.

Worse still are AI deployments rooted in distrust of employees. Some executives make these assumptions about workers:

They are lazy and need AI to monitor productivity.

They are incompetent and need AI to correct their mistakes.

They are redundant and need AI to replace them.

In reality, many employees are actually poorly equipped, poorly led, or stuck in broken processes. AI that surveils, second-guesses, or replaces them without addressing systemic issues will only breed resentment and inefficiency.

Before investing in AI, organizations must ask these questions:

1. What specific problem are we trying to solve? (Not just, “We need AI because everyone else has it.”)

2. Is this a problem of data, process, or culture? (AI can’t fix bad leadership or siloed workflows.)

3. Have we exhausted non-AI solutions? (Sometimes, better training or process redesign works better.)

4. Are we prepared to govern AI properly? (Without oversight, AI can amplify biases or errors.)

AI holds tremendous potential—but only when applied to well-defined problems. Right now, much of the AI frenzy resembles past tech bubbles, where hype outpaced practical utility. The organizations that will succeed with AI do the following:

Acknowledge their real inefficiencies (not just the ones vendors claim AI fixes).

Listen to employees about pain points before imposing AI solutions.

Treat AI as an enabler, not a magic wand

Otherwise, we’ll see more expensive experiments ending like the oil company’s KM project: two broken systems instead of one and the same old problems lingering underneath.

AI isn’t the problem. Not knowing why you’re using it is.

KMWorld Covers
Free
for qualified subscribers
Subscribe Now Current Issue Past Issues