70% of AI features fail to move revenue needles: SaaStr's brutal audit

The Gist

• 70% of 2025 AI features had zero measurable impact on core revenue metrics
• Only 10% drove actual revenue growth or retention improvements
• Valid AI features must enable 20-50% price increases or 20%+ retention bumps
• Most copilots are ghost towns despite saving users 10 minutes daily

• 70% of AI features fail to move revenue needles, with only 10% driving significant revenue impact.
• AI features must directly increase ACV, reduce churn, or expand NRR past 120% to be considered successful.
• Most AI copilots fail because they don't translate user time savings into CFO-relevant value.
• The market has moved past the phase of shipping AI features for novelty; results are now the only metric that matters.
• If removing an AI feature would have no impact on revenue in 90 days, it’s not a revenue-driving feature.
• Successful AI features are built by starting with revenue impact and working backward, not by focusing on engagement or demos.

• Evaluate AI features based on their direct impact on revenue metrics like ACV, churn, and NRR.
• Remove AI features that don’t materially impact revenue within 90 days.
• Focus on building AI features that solve CFO-relevant problems, not just user convenience.
• Start AI development with revenue impact as the primary success metric, not engagement or novelty.

• 70% (Percentage of AI features that fail to move revenue needles.)
• 10% (Percentage of AI features that drive significant revenue impact.)
• 120% (Target NRR expansion for AI features to be considered successful.)