Nearly every portfolio company conversation right now touches on AI in some form, and there’s a specific pattern worth naming honestly, because it’s shaping a lot of capital allocation decisions in ways that deserve more scrutiny than they’re currently getting. 

The pattern is a gap between AI initiatives that are genuinely producing measurable operational value and AI initiatives that are primarily producing a compelling narrative for the next board meeting or the next fundraise. 

Both categories exist in real numbers right now. The genuinely valuable initiatives tend to be narrow, specific, and boring to describe, a support ticket triage system that’s measurably reduced resolution time, a sales process step that’s genuinely faster with a specific tool. 

The narrative-driven initiatives tend to be broader, more ambitious in description, and considerably harder to point to a specific, measured outcome for, “we’re transforming our customer service function with AI” without a specific before-and-after metric attached. 

The reason this matters beyond intellectual honesty is that capital and management attention are being allocated toward AI initiatives right now partly on the strength of the narrative rather than purely on evidence of return, because the current environment rewards being seen to be doing something with AI, and that reward structure creates a genuine incentive to overstate what a specific initiative is actually delivering. 

The practical discipline worth applying to any AI initiative currently being discussed in a portfolio company: ask for the specific, measured before-and-after metric, not the description of the initiative’s ambition. If a genuinely specific metric exists and has moved, that’s a real result worth continuing to invest in. 

If the answer is a description of scope and potential rather than a measured outcome, that’s worth treating as an early-stage bet, not a proven value driver, regardless of how confidently it’s being presented.