Skip to content

AI-assisted engineering: How AI is transforming software development

A research-driven look at how AI coding tools are changing productivity, quality, and the developer experience

Taylor Bruneaux

Analyst

AI has changed how developers write, test, and ship software—but not always how organizations measure the results. Across 435 companies and 135,000 developers, DX’s research shows a clear pattern: success with AI coding tools isn’t determined by adoption alone, but by how systematically teams measure their impact.

AI-assisted engineering represents a fundamental shift in how software is created. The real value of AI coding assistants emerges across three dimensions—utilization, impact, and cost. Together, these dimensions form a complete view of how AI affects productivity, code quality, and developer experience.

The data is unambiguous. With 91 percent of engineering organizations now using AI tools, the conversation has moved beyond whether to adopt AI—it’s about how to measure what it’s actually doing. The difference between acceleration and inefficiency comes down to structure: enablement, workflow adaptation, and continuous feedback.

Organizations that treat AI-assisted engineering as measurable infrastructure—not an experiment—see faster delivery, improved developer satisfaction, and better engineering allocation. DX’s research and platform capabilities provide the clarity leaders need to quantify what’s working, identify friction, and optimize their investments.

This hub brings together DX’s latest research, frameworks, and real-world studies on AI-assisted engineering—clarifying not just how to use AI, but how to measure its true effect on engineering performance.

What is AI-assisted engineering?

AI-assisted engineering represents the integration of AI coding tools throughout the software development lifecycle to amplify developer effectiveness. As described in DX’s AI Measurement Framework, successful AI adoption requires understanding three dimensions: utilization patterns, productivity impact, and cost optimization.

With 91% adoption across engineering organizations, AI coding assistants have moved from experiment to essential infrastructure. But adoption doesn’t equal impact. Organizations see dramatically different outcomes based on how they approach enablement, measurement, and rollout strategy.

Why AI-assisted engineering is relevant today

AI tools save developers an average of 3.6 hours per week, enable 60% higher PR throughput for daily users, and cut onboarding time in half. But these gains aren’t automatic. They require new developer skills, adapted workflows, and comprehensive measurement to ensure speed doesn’t compromise quality.

For engineering leaders, AI-assisted engineering creates opportunities to accelerate delivery, improve developer satisfaction, and optimize engineering allocation. Organizations tracking AI impact through AI Usage Analytics and AI Impact Analysis gain visibility into what drives value and where investments pay off.

Research and evidence

DX has conducted the industry’s most comprehensive research on AI-assisted engineering, analyzing data from over 135,000 developers across 435 companies. This research reveals how AI tools actually impact productivity, quality, and developer experience—moving beyond vendor marketing claims to show what drives real value.

AI-Assisted Engineering: Q4 Impact Report

DX’s original research analyzing AI adoption and impact across 135,000+ developers and 435 companies. Key findings: 3.6 hours saved per week per developer, 22% of merged code AI-authored, 60% higher PR throughput for daily users. The research reveals that junior engineers adopt fastest (41.3% daily usage), staff+ engineers save the most time (4.4 hrs/week), and organizations with structured enablement see 8% better code maintainability and 19% less time loss.

Measuring AI Code Assistants and Agents

DX’s framework for tracking utilization, impact, and cost of AI coding tools. Developed in collaboration with leading organizations and AI vendors to provide vendor-agnostic measurement across three critical dimensions.

Microsoft’s 3-week Copilot study

DX analysis of early empirical research on GitHub Copilot productivity gains, highlighting why comprehensive measurement matters beyond vendor-provided metrics.

Impact of Copilot at ANZ Bank

DX case study examining enterprise-scale AI measurement approaches in regulated financial services context.

How DX measures the impact of AI-assisted engineering

The AI Measurement Framework

DX’s AI Measurement Framework provides research-based metrics across three dimensions that capture complete AI impact:

  • Utilization metrics reveal how developers actually use AI tools, not just whether they have access. Track AI tool usage through DAUs and WAUs, percentage of PRs that are AI-assisted, percentage of committed code that is AI-generated, and tasks assigned to autonomous agents.
  • Impact metrics capture whether AI improves engineering effectiveness. Measure AI-driven time savings, developer satisfaction, DX Core 4 metrics including PR throughput and code maintainability, and quality indicators like change confidence and change failure rates.
  • Cost metrics ensure AI investments deliver returns. Track AI spend per developer, net time gain (savings minus costs), and for autonomous agents, agent hourly rate comparing human-equivalent hours against AI spend.

Organizations implementing comprehensive measurement move past vendor metrics and understand what actually drives value. For complete measurement guidance, see the AI Measurement Hub.

DX platform capabilities for AI

DX operationalizes the AI Measurement Framework through integrated analytics that track AI usage, measure productivity impact, and connect AI adoption to business outcomes. These capabilities provide engineering leaders with visibility into what drives value from AI investments across their organization.

  • AI Usage Analytics tracks adoption patterns across tools and teams, revealing who uses AI, how frequently, and for which tasks.
  • AI Impact Analysis measures productivity effects and quality indicators, connecting AI usage to delivery velocity and developer experience outcomes.
  • AI Code Metrics provides visibility into AI-generated code distribution across your codebase, enabling quality and maintainability tracking.
  • Team Dashboards surface AI metrics alongside broader productivity indicators for comprehensive view of engineering effectiveness.

Proven techniques for effective AI-assisted engineering

Advanced prompting strategies

Research shows that AI-driven coding requires new techniques many developers don’t know yet. The gap between basic and advanced AI usage is dramatic. Developers who master sophisticated prompting see significantly better results.

DX’s Guide to AI-Assisted Engineering documents proven techniques from AI-savvy engineers, like:

  • Meta-prompting embeds instructions within prompts to guide how models approach tasks, reducing back-and-forth clarifications.
  • Prompt-chaining builds workflows where one prompt’s output becomes the next prompt’s input, enabling complex code generation from simple conversations.
  • Few-shot prompting provides examples of desired output, dramatically improving result quality and structure.
  • Multi-model evaluation leverages AI’s objectivity to compare solutions across models and identify superior approaches.
  • Multi-context inputs use voice and images alongside text, speeding interactions by 30% or more.

High-value use cases

Based on self-reported time savings from developers saving one or more hours weekly with AI, ten use cases deliver the most value: stack trace analysis, refactoring existing code, mid-loop code generation, test case generation, learning new techniques, complex query writing, code documentation, brainstorming and planning, initial feature scaffolding, and code explanation.

Each use case addresses specific friction points in the developer workflow. Implementation guidance with prompt examples and sample outputs is available in the complete guide.

Strategic implementation and enablement

The AI strategy playbook for engineering leaders

Rolling out AI tools successfully requires more than purchasing licenses. DX’s AI Strategy Playbook outlines how engineering leaders drive real impact:

Executive buy-in and evangelism drives adoption from the top. Leaders must showcase successful teams, remove barriers to usage, and maintain momentum through visible support.

Structured enablement programs make the critical difference. Organizations with better enablement see measurably better outcomes across code quality, developer confidence, and time savings.

Comprehensive measurement frameworks track adoption, impact, and cost together. Without measurement, organizations can’t determine whether AI investments deliver returns.

Quality guardrails adapt code review and testing processes for AI-generated code. Organizations should continue existing quality practices while developing new approaches for AI outputs.

Acceptable use policies balance security requirements with developer experimentation. Clear policies prevent shadow AI vulnerabilities while preserving innovation.

Communicating AI impact to stakeholders

Engineering leaders need to demonstrate ROI from AI investments with evidence, not anecdotes. Effective communication requires understanding what different stakeholders value.

Executives focus on business outcomes, ROI, and competitive advantage. Product leaders care about delivery velocity and time-to-market. Engineering managers prioritize developer satisfaction, quality metrics, and retention. Developers want practical benefits, reduced toil, and career development.

Each audience requires different framing of the same underlying data. Connect AI adoption patterns to outcomes each stakeholder values using Executive Reporting and Custom Reporting capabilities.

Tool selection and optimization

Comparing AI coding tools

The AI tool landscape evolves rapidly. Understanding trade-offs helps optimize selection and avoid vendor lock-in.

Organizations should evaluate tools based on capabilities, security, cost, and workflow integration. The performance landscape shifts rapidly, making vendor flexibility critical.

Real-world outcomes

Organizations maximizing AI impact

Workhuman increases AI ROI by 21% Through comprehensive measurement and optimization with DX, Workhuman demonstrated clear business value from AI tool investments and identified opportunities for further gains.

GitHub measures developer experience with DX Even the creators of Copilot use DX to understand how AI tools impact their own developers’ productivity and satisfaction.

Booking.com scales from under 10% to 70% adoption Structured enablement across 3,000+ developers drove industry-leading AI adoption rates and measurable productivity improvements.

Fintech achieves measurable Copilot gains Structured rollout approach with comprehensive measurement demonstrated clear productivity impact from GitHub Copilot deployment.

Implementation guidance

Copilot structured rollout best practices Proven approaches for enterprise AI tool deployment emphasizing enablement, measurement, and iterative optimization.

Explore AI measurement and optimization

AI Measurement Hub: Complete frameworks, metrics, and implementation guidance for measuring AI coding tools comprehensively.

Guide to AI-Assisted Engineering: Proven techniques and workflows for maximizing AI adoption and developer productivity.

AI Strategy Playbook: What engineering leaders need to know about driving AI impact across organizations.

Communicating AI Impact: How to measure and communicate AI-assisted engineering impact to different stakeholders.

Understand broader productivity context

Developer Experience Index (DXI): Research-based framework capturing the full range of factors that drive engineering effectiveness.

DX Core 4: Unified measurement framework balancing speed, effectiveness, quality, and business impact.

Engineering Productivity: Comprehensive approach to measuring and improving how engineering teams deliver value.

TrueThroughput: Gaming-resistant output measurement that accounts for work complexity and value, not just volume.

Compare tools and approaches

AI Engineer vs Software Engineer: Understanding evolving role definitions as AI transforms software development.

AI is the future, but productivity matters more than ever: Why core productivity fundamentals remain critical even as AI adoption accelerates.

See how DX measures AI impact across your organization

The most successful organizations don’t just deploy AI tools. They measure adoption patterns comprehensively, connect AI usage to productivity outcomes, and optimize based on data. They understand that AI is an accelerant that amplifies existing practices, making strong measurement foundations essential.

DX provides the industry’s most comprehensive platform for measuring AI-assisted engineering impact. Track utilization across all major tools, measure effects on productivity and quality, and demonstrate ROI to stakeholders.

Request a demo to see how DX helps engineering leaders extract real value from AI investments.

Published
November 12, 2025