5 Questions to Answer Before Investing in Your Next AI Project

1) What business problem are we actually trying to solve?

• Don't start with "Where can we use AI?" Instead, start with a measurable business problem; margin leakage, excess inventory, downtime, labor productivity, forecasting accuracy, customer retention, etc. If the problem isn't clear, the AI project won't be either.

2) What would solving this

problem be worth?

• Quantify the opportunity before investing. What happens to revenue, margin, cost, working capital, productivity or risk if the problem improves? This creates the economic case and gives you something against which to measure the investment.

  
  

3) Do we have the data needed to solve it?

• The relevant information may exist across ERP, CRM, MES, financial, HR, POS, spreadsheets and other systems. The question isn't simply whether you have data, it's whether the necessary data is accessible, connected, reliable and understood well enough tot support the use case.

  
  

4) Is AI actually the right solution?

• Some problems need AI, other need a better process, clearer accountability, automation, conventional analytics or simply a good dashboard. Using AI where a simpler solution would work isn't innovation, it's unneccasry cost and complexity.

  
  

5) How will we know if it worked?

• Define the expected business outcome / what success looks like before implementation. What metric should move? By how much? Over what period of time? Who owns the result? The project isn't successful because the technology works, it succeeds when the business result improves.

The best AI strategy isn't the one with the most AI projects, it's the one that invest in the right ones.