Over the past 6–9 months, our team has been working on one of the most ambitious research efforts I’ve led at Forrester: a dual-track program combining dozens of qualitative interviews with development leaders and practitioners and a rigorous quantitative analysis of AI’s employment impact across the software development lifecycle (SDLC). Working shoulder-to-shoulder with my colleagues Michael O’Grady and Lorenzo Annachiarico, we set out to go beyond the hype and produce a clear, evidence-based view of how predictive, generative, and agentic AI are reshaping software work — and what leaders should do now.

AI Doesn’t Replace The Development Workforce — It Changes It

Our new trends report, “AI Is Evolving The Development Workforce In Dramatic Ways,” captures an inflection point: AI isn’t replacing developers; it’s changing what developers (and their teammates) do day to day. As AI takes on more of the artifact creation (code, tests, docs), human roles expand toward orchestration, systems thinking, governance, and business alignment — the work of steering AI and stitching value across the SDLC.

This rebalancing comes with team-level consequences. We see role convergence (classic front-end/back-end/QA boundaries blur), roles evolving, smaller cross‑functional pods, and rising demand for T‑ and E‑shaped skill profiles that mix coding, product, data, and governance literacy. A big new topic is emerging: While vibe coding is attractive, vibe engineering (or context engineering) will be the focus for software development in 2026. The flipside: new challenges in training juniors when entry‑level tasks are automated and in building organizational trust for AI systems that can still err.

AI Changes Work Volume And Mix Across Development

Our companion report, “The Quantitative Employment Impact Of AI On The Software Development Lifecycle,” quantifies where AI is changing work volume and mix across requirements, design, build, test, release, and run. The headline: predictive, generative, and agentic AI are catalyzing the most dramatic workforce shift the SDLC has seen in years — with hours moving from repetitive artifact production toward higher‑leverage activities, such as workflow orchestration, architecture validation, controls, and customer value realization. Leaders who plan for transformation rather than reduction will outperform competitors. Our study highlights the roles (or tasks in the roles) that will be displaced or augmented by AI.

Context helps — in earlier Forrester forecasting work, we found that generative AI will influence far more jobs than it eliminates by 2030. This pattern is already playing out in software: AI changes the contour of roles and skills more than it erases them. The strategic question shifts from “What jobs go away?” to “How do we redesign jobs, teams, and career paths to harness AI responsibly?”

Five Leadership Takeaways

  1. Redesign roles for orchestration, not just production. Treat developers, SDETs, product managers, and platform engineers as AI conductors who guide agents, compose workflows, and enforce guardrails. This demands explicit skills in prompt/constraint design, evaluation, and governance — not just language/framework expertise.
  2. Protect (and modernize) the junior pipeline. If AI absorbs many entry‑level tasks, you’ll need intentional apprenticeship models: pair juniors with seniors; rotate them through evaluation, observability, and safety work; and create “practice dojos” where they learn to critique and improve AI output. Our research highlights this as a looming risk area for talent sustainability.
  3. Reskill — don’t downsize. The fastest ROI comes from upskilling existing teams to apply AI across the whole SDLC (not just coding). Shift measurement from activity metrics to outcomes, such as customer value, cycle time to impact, reliability, and risk posture. The right playbook is to hold onto developers and boost them with AI — not cut headcount.
  4. Institutionalize governance-by-design principles. As agentic patterns spread, treat “governance as code” and “observability as code” as first‑class artifacts. Invest in evaluation harnesses, lineage, and policy enforcement so teams can scale AI safely and repeatedly — a pattern we also see gaining traction across the industry as AI use matures.
  5. Show the money — across top and bottom lines. Our broader pulse data indicates enterprises adopting generative AI already report improvements in employee productivity, CX, and revenue — but leaders still under-measure value when they fixate on activity metrics. Tie AI to auditable outcomes to win continued investment.

The Bottom Line

Software development is moving from “people producing artifacts, assisted by tools” to “teams orchestrating AI‑accelerated systems with human judgment at the core.” That shift won’t happen by accident. It requires redesigned roles, outcome‑based metrics, a protected talent ladder, and governance that earns trust. Leaders who act now will capture compounding returns as AI reshapes not just coding, but the entire path from idea to customer value.

To find out more, read the two companion documents: 1) AI Is Evolving The Development Workforce In Dramatic Ways and 2) The Quantitative Employment Impact Of AI On The Software Development Lifecycle. You can also reach out for a guidance session or an inquiry. If you aren’t a client and have any comments, suggestions, or want to share your successful story or insurmountable challenges you’ve faced, email me directly at dlogiudice@forrester.com.