Boards want 10x, and most companies are not measuring
The 10x number keeps showing up in board decks while few companies measure anything. The pressure rolls downhill, and the gap between what gets reported and what happens keeps widening.
Written by
-
Benedikt Stemmildt Founder & Co-CEO
Published: 5 May 2026 Updated: 2 min read
The 10x number keeps showing up in board decks. I get why: it is the kind of headline that makes a budget feel like it is already working.
In most of those companies, nobody is actually measuring.
Felt efficiency, no number behind it
A Dataiku survey of 600 CIOs, run by The Harris Poll, puts it cleanly: 98% report increased board pressure to prove AI ROI. Most of them have nothing on either side of the equation.
So the pressure rolls downhill. Board to CEO to CIO to CTO to engineering. Show the 10x we already told ourselves is happening.
What the measured numbers say
Meanwhile the actual numbers stay in the same band. In PwC’s 2026 CEO survey, 56% of CEOs say they have seen no significant financial benefit from AI so far. MIT NANDA found that 95% of organizations are getting zero return on GenAI. Laura Tacho’s DX data on 121,000 developers across 450 companies puts the productivity gain around 10%.
Tacho herself is pretty clear about why. Adoption is not the input variable that moves productivity. Enablement is. And enablement means more than training: the processes around the work, the review queues, what teams have room to change about how they ship.
DORA made the same point in its 2025 report. AI does not fix a team, it amplifies what is already there. Hand a slow review process AI and you get faster code piling up at a slow review queue.
METR’s randomized controlled trial made the perception gap concrete. Experienced open-source developers were 19% slower with AI tools. They believed they were 20% faster. The board only ever sees the felt number.
What the telemetry shows
Faros telemetry from over 10,000 developers across 1,255 teams (opens in a new tab) finds that developers on teams with high AI adoption touch 47% more pull requests per day.
PR-based metrics are output, not outcome. I think we will get to better numbers over time. But the best measure a team has today beats waiting for the perfect one.
The two questions to ask back
So when the board asks for the 10x, the questions I would ask back are: are we measuring at all? And did anything around the work actually change before we counted?
If the answers are no and nothing, you measured exactly what Tacho’s sample captured: adoption without enablement, at 121,000-developer scale.
What did you measure? And when?
Further reading
Written by
-
Benedikt Stemmildt · Founder & Co-CEO
More on Measuring AI impact
- AI pays off in maintenance work, not in the demos DX Research ranked ten developer tasks by AI time savings. The tasks most rollouts start with, code explanation and scaffolding, sit at the bottom, and the gains land in brownfield work.
- Clean codebases decide who gets value from AI Stanford research on 120,000 developers, presented in a talk, found codebase cleanliness predicted AI value better than token spend. Teams with weak codebases regress past a certain spend.
- Measure AI impact, not AI usage Most companies track how much their developers use AI tools. The DX AI Measurement Framework adds impact and cost, and the metrics that survive measure what reaches production.
Test the constraint in your system
The Factory Readiness Assessment connects this research to your value stream, your data, and the next decision.
Training on this topic: Agentic Engineering for Engineering and Product LeadersWant to check one team first? The self-check: twelve yes or no questions, three minutes, instant result, no account. Take the self-check