Software development and application integration are no longer episodic activities. Driven by evolving business models, regulatory obligations, platform modernisation, cybersecurity and rapid AI adoption, technology delivery is becoming continuous. At the same time, Agentic AI is beginning to reshape engineering, delivery and operations.
Delivery was designed for human pace
Traditional delivery models — whether waterfall, Agile, or scaled frameworks such as SAFe — were designed around human-paced execution. Planning, sprint rhythms, governance, estimation and coordination all assume bounded cycles and sequential flow. An AI-native SDLC operates differently.
A co-creation space between humans and AI
In an AI-enabled engineering environment, software delivery increasingly becomes a co-creation space between humans and AI agents. Agents generate options, write code, test, detect risks and recommend pathways, while humans focus on context, judgement, validation and accountability. Design, build, test and deploy compress into continuous cycles. Application integration accelerates. Feedback loops shorten dramatically. A seamless continuum.
Large enterprises are already experimenting with autonomous engineering agents. Goldman Sachs, for example, is piloting Devin as part of a hybrid human–AI engineering model — an early shift from AI assistance toward AI participation in software delivery.
If software development fundamentally changes, can delivery management remain unchanged?
The assumptions already shifting
Some delivery assumptions already appear to be moving:
• Scrum ceremonies increasingly coordinate people and agents.
• Backlog refinement becomes intent refinement.
• Velocity gives way to measures of outcome quality, verification, resilience and business fit.
• Retrospectives shift toward continuous telemetry and adaptation.
AI agents can increasingly trace dependencies across applications, programs, regulatory obligations and business priorities. Sequencing becomes dynamic. Prioritisation becomes scenario-driven and near real-time. Over time, the boundaries between project, program and portfolio management may begin to blur — as strategic alignment, governance and benefits realisation become integrated and data-driven, assisted by AI agents at every level.
From socio-technical to complex adaptive
Enterprise delivery has always been a socio-technical system, shaped by interactions between people, process, technology and governance. AI changes this environment by increasing speed and autonomy while complicating context tracing. As human and AI work overlap, new tensions emerge around accountability, transparency, governance and trust. Enterprise delivery increasingly behaves like a complex adaptive system, where traditional control models struggle to keep pace with distributed autonomy and compressed decision cycles.
The real challenge
The challenge is no longer simply automating delivery with AI. It is redesigning the enterprise delivery ecosystem for an AI-native operating model.
To navigate this shift, I’m exploring an interpretive framework — SEACO™ (Stewarded, Ethical, Adaptive Co-Orchestration) — to help enterprises rethink delivery in environments where humans and AI agents co-create outcomes.
Emerging research is beginning to question whether delivery models designed for human-paced software engineering remain fit for purpose in an AI-native SDLC. If software engineering changes, then project, program and portfolio management may need to evolve with it.
References
Müller, R., Locatelli, G., Holzmann, V., Nilsson, M. and Sagay, T. (2024) ‘Artificial intelligence and project management: empirical overview, state of the art, and guidelines for future research’, Project Management Journal, 55(1), pp. 9–15.
Banh, L., Holldack, F. and Strobel, G. (2025) ‘Copiloting the future: how generative AI transforms software engineering’, Information and Software Technology, 183, p. 107751.
Assalaarachchi, L.I., Masood, Z., Hoda, R. and Grundy, J. (2026) ‘Toward agentic software project management: a vision and roadmap’, arXiv preprint arXiv:2601.16392. Available at: https://arxiv.org/abs/2601.16392 (Accessed: 24 May 2026).
McKinsey & Company (2026) ‘The AI revolution in software development’. Available at: https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-ai-revolution-in-software-development (Accessed: 24 May 2026).
IBM (2026) ‘Goldman Sachs’ first AI employee, Devin’. Available at: https://www.ibm.com/think/news/goldman-sachs-first-ai-employee-devin
I write about how enterprise technology transformation must evolve in the agentic era — including SEACO™, a socio-technical model for embedding AI into delivery with governance designed in. More at orchanex.ai.