Talk to us
2027 Edition — Available Now

IT Outlook
Report 2027

The trends and insights IT leaders need to govern what AI is building in 2027.

Download free report
1.4% of enterprise code in production in the IT industry is AI-generated
91% of IT-industry code falls below our recommended maintainability rating
€870K saved in developer time, per system, per year, by cutting code-level technical debt
IT Outlook Report 2027 cover mockup

Trusted by

GitLab Keylane Centric Software Progress TerraQuest bynder
The 2027 pattern

AI accelerates whatever discipline already exists.

90% of technology professionals now use AI at work, and the volume of code entering organizations is rising faster than at any point in software history. Review capacity hasn't kept up. Where quality is measured and managed, AI accelerates delivery. Where it isn't, AI accelerates debt, cost, and exposure at the same speed. Six chapters break down where that pattern is already showing up.

Chapter 01 — AI coding

AI-generated code carries roughly double the security risk of human-written code

It's also less maintainable than human-written code, especially in large systems, and non-agentic AI spend for a 50-developer team already costs the equivalent of nearly one extra developer per year.

Chapter 02 — Technical debt

€870,000 in developer time, per system, per year

That's what's recovered by moving a system from 2-star to 4-star maintainability. Systems with lower code-level technical debt also see up to 72% stronger security compliance.

Chapter 03 — Maintainability

91% of IT-industry code sits below the recommended threshold

That's above the 86% global average. The average portfolio also runs 3 to 4 core technologies, adding complexity most teams haven't accounted for.

Chapter 04 — Security

67% of IT-industry code has a low degree of security controls

Large systems, the ones most likely to be business-critical, score lowest of all. The average system carries 20 critical security findings.

Chapter 05 — Architecture

39.7% of systems fall below our architecture quality rating

Second-worst of any industry benchmarked. Nearly 60% of AI leaders cite legacy-system integration as their top barrier to deploying agentic AI.

Chapter 06 — AI systems engineering

72% of AI systems score below the maintainability threshold

Only 1.5% of enterprise systems in production currently qualify as AI systems, but human-in-the-loop builds already score far higher than fully autonomous ones.

“

The findings on the pages that follow come from our own benchmark data on IT and software services systems specifically, set against the best independent research available. Some of it is encouraging. Much of it should be food for thought. All of it is meant to help you go faster — responsibly.

Luc Brandts, CEO of Software Improvement Group
Luc Brandts CEO, Software Improvement Group
Who it's for

Built for the leaders governing software in the AI era.

Each chapter speaks to a different seat at the table. Start with yours, then read the rest to understand what your peers are working against.

CIOs & IT leaders

See where budget is locked up

Technical debt already consumes 21–40% of total IT spend. Find out where that capacity sits, and what recovering it is worth.

Chapters 02, 03
CTOs & engineering leaders

Govern AI-assisted development

Understand what agentic coding actually costs in tokens, and where AI output needs a harness before it reaches production.

Chapters 01, 06
CISOs & security leaders

Close the AI security gap

See where AI-generated and open-source code is carrying more risk than your review process is built to catch.

Chapter 04
Architecture & delivery leaders

Fix what's blocking agentic AI

Legacy-system integration is already the top barrier to agentic AI adoption. See what that means for your roadmap.

Chapter 05
Free download

Get the IT Outlook Report 2027

  • Six in-depth chapters covering AI coding, technical debt, maintainability, security, architecture, and AI systems engineering
  • Benchmark data drawn from more than 30,000 systems and over 400 billion lines of code
  • A sector-specific lens on IT and software services, set against the best independent research available
  • Real experiments comparing AI-only and human-in-the-loop builds, and what guardrails do to agentic output

This field is for validation purposes and should be left unchanged.
Name*
Privacy*

Read our methodology

This field is for validation purposes and should be left unchanged.
Name*
Privacy*

Watch Core in action

This field is for validation purposes and should be left unchanged.
Name*
Privacy*

Watch Axis in action

This field is for validation purposes and should be left unchanged.
Name*
Privacy*

This field is for validation purposes and should be left unchanged.
Name*
Privacy*

This field is for validation purposes and should be left unchanged.
Name*
What type of partnership are you interested in?*
Privacy*

Register for access to Summer Sessions

This field is for validation purposes and should be left unchanged.
Name*
Privacy*