JOSEPH GABRIEL LAGONSIN
Information Editor
Safe Code Warrior has launched the SCW AI Adoption Mannequin for software program improvement groups, a framework aimed toward Chief Info Safety Officers managing the shift in the direction of extra autonomous AI use.
The mannequin outlines three phases of AI use in improvement: AI-Assisted, AI Native and Agentic. Every stage is tied to completely different danger ranges, developer ability necessities and governance controls as organisations deliver AI instruments into coding, no-code workflows and different software program creation processes.
The framework is meant to assist safety leaders assess the place their organisations sit on the AI adoption curve and determine what coaching and oversight ought to observe. It comes as firms face rising stress to adapt safety practices to improvement environments the place AI methods can generate, refine or orchestrate code with much less direct human enter.
Trade analysts have additionally pointed to the tempo of change. Gartner’s 2026 Hype Cycle for Safe Software program Engineering mentioned AI-augmented improvement is increasing the assault floor sooner than conventional controls can scale, whereas growing the significance of safe coding abilities.
Three phases
Beneath the mannequin, AI-Assisted describes improvement work during which AI instruments assist programmers however don’t dominate the workflow. AI Native refers to extra built-in use of AI throughout improvement duties, whereas Agentic describes a stage during which autonomous methods tackle a broader orchestration position within the improvement lifecycle.
The construction is designed to assist organisations join AI use with software program danger indicators and sensible governance measures. It is usually supposed to assist the transition from the Software program Improvement Lifecycle to what Safe Code Warrior calls the Agentic Improvement Lifecycle.
The announcement displays a broader change in who contributes to software program danger inside firms. AI use is not confined to specialist engineering groups, with non-developers more and more utilizing no-code instruments and what Safe Code Warrior described as vibe coding approaches to construct purposes or automate workflows.
That enlargement has difficult the duty for safety groups. Conventional controls, typically constructed round standard software program engineering groups and established evaluation processes, are being examined by sooner and extra extensively distributed types of software program creation.
Governance focus
The framework offers organisations a strategy to determine their present stage of AI adoption, map related coaching to builders and different customers, and put governance controls in place earlier than AI use turns into extra autonomous. It may additionally assist safety leaders reveal returns from governance and coaching by linking behaviour modifications to danger discount.
Safe Code Warrior argued that coaching stays central as AI instruments grow to be extra frequent in improvement work. Fairly than relying solely on technical controls to detect errors in AI-generated code, organisations want builders and different software program creators who can use these instruments safely from the outset, it mentioned.
The argument can be tied to price in addition to safety. In response to Safe Code Warrior, Gartner has predicted that by 2027 greater than 40% of agentic AI tasks will likely be deserted due to uncontrolled prices and poor danger controls.
Pieter Danhieux, Co-founder & Chief Government Officer at Safe Code Warrior, mentioned the rise of AI-assisted improvement is altering the position of builders and the expectations positioned on them.
“In our present AI-powered improvement, writing strains of code is sort of free, however builders are nonetheless on the hook for safe outcomes. Their safety abilities must evolve from code author to creator & orchestrator,” mentioned Danhieux.
He mentioned the corporate constructed the framework in response to a necessity for governance approaches that match newer improvement strategies.
“CISOs want an strategy to ADLC governance that’s as fashionable because the methodology itself, one which follows an adoption mannequin designed for agentic AI’s evolving, adaptive strategy to software program improvement. We have constructed this framework to assist organizations flip safe AI adoption and AI governance from a reactive train right into a measurable, scalable self-discipline,” mentioned Danhieux.