5 Key Indicators That AI Isn't The Answer
Sometimes it takes greater leadership to “not” do a thing than it does to succumb to the pressure of going along with the crowd. This is just as true in business as it is in high school. Most (all ?) leaders want to be seen as innovative and technologically savvy. This, combined with the pressure to “not be left” behind the AI wave (craze ?), creates a compelling urge to implement AI for something in your business. And that is a recipe for disaster.
This post is not about the AI bubble about to burst, which is one of the two narratives you most read in the blog-o-sphere. This post provides a reality check, or strategic reasons if you like, for your decision to pursue AI to help your business (e.g., drive business value). Pursuing AI not only requires significant organizational change, but also requires adoption of new and rapidly evolving technologies. In many cases, the best ROI is to make an informed decision not to adopt a new tool.
There are 5 key considerations that should weigh in on the decision to pursue AI within your business or business areas.
Clear Rules Exist
The flow from customer request to delivery (the Value Stream) follows a clearly defined and easy to codify set of rules (e.g., a deterministic process). These kinds of processes are easy to understand, easy to codify using rule-based automation and procedural scripts. More importantly, rule-based automation provides lower cost, higher compliance, and easier audits.
2. Data Volume and Variability
AI systems rely on generalizing from examples found in your data. Small datasets mean fewer examples to learn from. This leads to models that are inflexible and more likely to fail. The reality is, if you have small, mostly homogenous datasets, simple statistical models and heuristics will meet your business needs. This approach is significantly more cost-effective and yields (in most cases) recommendations and insights that are easier to understand.
3. Significant Regulatory Requirements
In highly regulated environments, such as finance or healthcare, organizations must explain and reproduce every decision they make. A black-box AI system that can’t justify its outputs introduces compliance and reputational risk. In these contexts, rule-based automation with human review offers both transparency and traceability, meeting regulatory demands while maintaining control.
The investment in ensuring the explainability (which drives traceability) in this kind of regulation is significant. Making ROI on implementing AI harder. It may be better (and more cost effective) to use rules in combination with a human in the loop automation to start.
4. Static Environments
This is the “If it ain’t broke, don’t fix it” view. In most cases, we argue against that sort of old school sentiment. That said, if the processes that support your Value Stream don’t change often, static systems or dashboards usually outperform adaptive models that require investment to maintain.
5. Unclear ROI
If you can’t clearly and easily link the AI investment to measurable financial outcomes, whether through revenue growth, cost reduction, or efficiency, pause the project. A disciplined AI strategy requires quantifiable ROI, not aspirational metrics. Without that clarity, you’re investing in experimentation, not transformation.
The statement above, “clearly and easily link the investment in AI to ROI” is not a throwaway sentence. You should be able to visualize (clearly) without significant effort that the investment in AI ties to greater sales, cost savings through growth with the same staff, more efficient operations driving staffing costs lower, etc.
The smartest AI strategy begins with knowing when not to use it. Technology is a multiplier—not a replacement—for disciplined thinking, well-designed processes, and human judgment. Using AI without clear operational problems to solve often results in AI theater, projects that impress but deliver no measurable value. Focus on improving data quality, automation maturity, and human decision systems before starting your AI journey. AI should be the last tool added, not the first.
