BOSTON, July 8, 2025 — Software program engineering platform Jellyfish printed the 2025 State of Engineering Administration report, which measures the impression of AI coding instruments on engineering work — from day-to-day workflows to large-scale organizational priorities. The sixth annual report gives an replace on how engineering groups are adopting AI, which coding instruments they’re deploying, how they’re paying for his or her AI investments, amongst different findings.
In response to the analysis, engineering groups are reaching substantial will increase in developer velocity and productiveness by way of AI coding. Software program engineers are united in a single space — the embrace of AI coding instruments — however cut up on the place the price range for these instruments ought to come from and the way AI stands to impression their productiveness shifting ahead. Notably, 45% of respondents anticipate offshoring as a proportion of their engineering headcount to extend over the subsequent 12 months.
Findings from the report embrace:
Adoption of and belief in AI has spiked: Total, 90% of groups are embracing AI coding instruments to enhance their engineering practices, a major improve from final 12 months when 61% of respondents mentioned their engineering organizations had embraced AI. Solely 3% aren’t utilizing AI instruments with no plans to take action.
AI is driving actual impression: Six in 10 (62%) respondents imagine they’re reaching a minimum of a 25% improve in developer velocity and productiveness by way of AI coding. Lower than 1% of respondents imagine AI is slowing them down.
In terms of AI, code writing is main the cost: The most well-liked use case for AI coding instruments is writing code. GitHub Copilot is the software of selection for 42% of respondents, adopted by Google’s Gemini Code Help, Amazon Q and Cursor.
The longer term will probably be a hybrid of people and AI: Lengthy-term, respondents count on a good portion of software program improvement work to shift from individuals to AI. Eight in 10 (81.4%) respondents imagine that a minimum of 25% of the engineering work people do right now will probably be dealt with by AI 5 years from now.
Corporations proceed to spend money on engineering: Six in 10 (61%) respondents mentioned their engineering price range elevated as a proportion of firm income from 2024 to 2025, in comparison with 19% who mentioned it remained the identical and 17% who noticed a lower.
Jellyfish surveyed 645 full-time professionals globally, in numerous engineering roles, together with particular person contributors, managers, and executives. Respondents — representing firms starting from small engineering groups with fewer than 10 individuals to enterprises with greater than 500 engineers — gave their ideas on AI, engineering funding, and metrics.
“Adopting AI in software program engineering is not non-obligatory, however unlocking its full worth requires greater than entry to coding instruments,” mentioned Andrew Lau, CEO and co-founder of Jellyfish. “It requires intentional measurement, structured enablement, and cultural funding. This isn’t a tooling improve — it’s an organizational transformation.”
