Data to cut through the hype: AI impact whitepaper

Building on previous research about what blocks teams from taking action on data, our newest research focuses on the impact of AI on developer productivity, codebase quality, and wellbeing.

Our research findings show:

• what matters for increasing adoption
• how AI impacts not only productivity but also hours worked
• what the best teams do differently to protect codebase quality while rolling out AI

Unlock the whitepaper for full details.

‍

More about the research:

From June-October 2025, we followed 500+ developers as their organizations rolled out new AI coding assistants. We combined telemetry data (about AI usage and developer workflows) with survey data and one-to-one interviews to get the full picture of AI impact. The organizations participating were a mix of Multitudes customers and non-customers.

The motivation for this research was to cut through the hype to see the real impact of AI on developer productivity. Existing research on AI and developer productivity has found a wide range of outcomes:

• A 2023 GitHub Copilot randomized controlled trial said that Copilot makes writing code 55.8% faster
• The 2025 METR study found that AI made developers 19% slower – and showed that people were almost exactly wrong when asked to self-report on AI’s productivity impact, since they thought they were 20% faster
• And we’ve seen every study in between – from the AI competence penalty slowing down adoption to an increase in copy-pasted code

We were also curious about what works or not in AI rollouts – since 2025 MIT research showed that 95% of AI initiatives at companies fail. 

Unlock the whitepaper to see what we found.

‍