AI analytics gains limited by hidden costs and messy data

Companies rolling out AI analytics run into three big costs: paying for the AI model itself, verifying its results, and making sure it has enough business context. Domo’s Ben Schein points out that measuring AI's productivity isn’t just about cheap models—reviewing and contextualizing output adds up. And while perfect data isn’t required to start, organizations do better choosing projects where data is “good enough” and results are clear.

• AI models' cheap up front but costly to verify
• Getting business context right adds effort
• Overly complex models can waste money
• You don't need perfect data to begin
• Pick use cases with clear, measurable results