In 2026 Engineering Leaders Are Shifting from AI Adoption to AI Accountability

For the past two years, the conversation around AI in software development has centered on one question: How do we get engineers to use AI?

In 2026, that question has changed.

Engineering organizations have largely moved beyond experimentation and pilot programs. AI coding tools are now embedded in daily workflows across the industry. The new challenge isn’t adoption – it’s proving value.

The latest State of Engineering Management report from Jellyfish reveals a clear trend: engineering leaders are increasingly focused on measuring AI’s return on investment, understanding rising token costs, and connecting AI-driven productivity gains to business outcomes.

AI is Delivering Measurable Productivity Gains

The good news for engineering leaders is that AI is producing tangible results.

According to the report, 64% of engineering professionals believe they are achieving at least a 25% increase in developer velocity and productivity through AI. Nearly seven in ten expect those gains to improve even further in the coming year. Meanwhile, only four respondents reported that AI was slowing them down.

The impact becomes even more pronounced among organizations that have embraced AI at scale. Respondents with very high levels of AI adoption were dramatically more likely to report improvements in productivity, efficiency, and job satisfaction than those with low adoption. Nearly all high-adoption organizations said AI positively influenced team productivity.

Jellyfish’s own platform data tells a similar story. Organizations with the highest AI adoption levels are merging roughly twice as many pull requests as low-adoption organizations, while maintaining code quality. Top adopters are even seeing autonomous agents contribute a meaningful percentage of production code.

The productivity story is becoming difficult to ignore.

The Next Challenge: Proving ROI

Despite widespread enthusiasm, a significant measurement gap remains.

Engineering productivity has become a board-level concern. Eighty-four percent of respondents say productivity is a top management priority, while 75% believe it is a strategic business concern. At the same time, only 46% of organizations are actively tracking AI-specific metrics such as adoption rates, acceptance rates, and model usage.

This disconnect creates a problem for leaders seeking to justify growing AI investments. It’s one thing to believe developers are moving faster. It’s another to demonstrate exactly how AI spending translates into business value.

As AI tools become more deeply embedded in software development, executive teams and boards are increasingly asking tougher questions:

• How much are we spending on AI?
• Which teams are generating the most value?
• Are productivity gains offsetting costs?
• What business outcomes are we achieving?

The organizations that can answer these questions with data will be far better positioned than those relying on anecdotes.

Token Costs are Becoming a Leadership Concern

Perhaps the most interesting shift in the report is the growing focus on AI spend.

When respondents were asked about AI adoption challenges, the number one concern wasn’t security, quality, or training – it was cost. Forty-two percent cited increasing AI tool expenses as a primary challenge, making it the most commonly reported obstacle to adoption.

In particular, engineering leaders are beginning to pay closer attention to token consumption.

Throughout 2025, many organizations prioritized experimentation. Developers were encouraged to maximize usage, test workflows, and discover what was possible. Cost management often took a back seat to learning.

That dynamic is changing.

As newer, more capable models drive both increased adoption and increased consumption, leaders are facing a new reality: AI spending can scale very quickly.

One platform leader quoted in the report described how widespread adoption of a coding assistant led to immediate productivity gains – followed by leadership concerns over unexpectedly high token costs. Access was ultimately reduced despite strong developer enthusiasm.

Stories like this highlight a broader challenge. Engineering organizations can no longer evaluate AI solely through the lens of productivity. They must also understand the economics behind that productivity.

Measuring Productivity and Spend Together

The report suggests that AI ROI will increasingly depend on combining usage data with engineering performance metrics.

Tracking token consumption alone doesn’t reveal whether AI investments are paying off. Likewise, measuring output without understanding associated costs provides only half the picture. Organizations need visibility into both sides of the equation:

• AI adoption and usage
• Token consumption and spend
• Developer productivity improvements
• Delivery outcomes
• Business impact

This is where platforms like Jellyfish offer an advantage. According to survey data, organizations using a Developer Productivity Insight platform were significantly more likely to track AI metrics, AI spend analytics, and AI-driven project planning than organizations without one. They were also more likely to report confidence in their productivity measurement strategies.

As AI investments grow, the ability to connect engineering activity directly to business outcomes may become one of the most important capabilities for technology leaders.

The Companies Pulling Ahead

The report’s most important takeaway is that AI adoption alone is no longer a competitive advantage. The leaders are increasingly separating themselves based on how effectively they measure, manage, and optimize their investments.

Organizations with the highest AI adoption report stronger productivity gains, greater efficiency improvements, and a more optimistic business outlook than their peers. They are also more likely to believe AI frees engineers to focus on higher-value roadmap work instead of maintenance and operational tasks.

The next phase of AI transformation won’t be defined by who has access to the best models. It will be defined by who can translate AI usage into measurable business results. In other words, the industry is entering a new era – one where AI adoption is expected, but AI accountability becomes the real differentiator.

For engineering leaders, that means the future belongs not just to the teams generating more code, but to the teams that can clearly demonstrate the return on every token they spend.

To learn more and read this year’s full State of Engineering Management report, you can download it here.