Enterprise AI Analysis
Toward Agentic Software Engineering Beyond Code: Framing Vision, Values, and Vocabulary
This paper argues for expanding the emerging vision of agentic Software Engineering (SE) beyond mere coding acceleration. It advocates for a 'whole of process' approach, grounding it in SE foundations and emerging agentic SE frameworks, while proposing preliminary values, principles, and well-defined vocabulary to guide the community towards a comprehensive process-level paradigm shift.
Key Strategic Outcomes
Transitioning to agentic SE offers profound opportunities to enhance efficiency, accelerate innovation, and ensure ethical development across the entire software lifecycle.
0%
Project Efficiency Increase
0%
Time-to-Market Reduction
0x
Innovation Velocity Uplift
0%
Ethical Compliance Rate
Deep Analysis & Enterprise Applications
Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.
SE Evolution Timeline
Whole of Process Vision
CRAFT Values & Principles
Vocabulary Guidance
The Evolution of Software Engineering
Software Engineering has undergone several paradigm shifts, each expanding its scope and methodology. From early informal practices to structured, iterative, agile, and augmented approaches, the field is now poised for an agentic transformation. This timeline highlights key eras:
Enterprise Process Flow
No Formal SE (1950s-60s)
→
Traditional SE (1970)
→
Iterative SE (1980s-90s)
→
Agile SE (2000s)
→
Augmented SE (2020s)
→
Agentic SE (Mid 2020s)
Expanding Agentic SE to a 'Whole of Process'
The paper proposes extending agentic SE beyond coding to encompass the entire software development lifecycle, integrating human and AI collaboration across various socio-technical aspects. This includes:
- Agentic SE Ethical Alignment: Integrating ethics-by-design, governance, and societal impact assessment for AI-driven processes.
- Agentic Requirements Engineering: Using AI to assist with customer, user, and subject matter expert elicitation, as well as cascading ethical requirements.
- Agentic Design: AI support for architecture, UI/UX, and design considerations for human benefit and equity.
- Agentic Development: AI-driven code design, production, testing, review, and repair, leveraging existing and emerging agentic AI tools.
- Agentic Operations: AI assistance in deployment, maintenance, and DevOps/DevSecOps activities.
This holistic approach ensures that agentic AI acts as a true paradigm shift, not just a coding accelerator, enhancing human capabilities and improving overall outcomes.
CRAFT Values and Principles for Agentic SE
To ensure a systematic and principled approach to Agentic SE, the paper outlines five core values and associated principles:
| Value | Principle | For Agentic SE |
|---|---|---|
| Comprehensive | Whole of Process Approach | Agentic SE needs to take a 'whole of process' approach to examining traditional SE areas and associated activities, actors, artefacts (AAA) and consider new ones to enable emerging human-agent ecosystems. |
| Human, Agentic, and Socio-technical Aspects | Agentic SE needs to include a focus on human, agentic, and socio-technical aspects of SE including roles played by human and agent actors, human and agentic aspects (e.g., human values, empathy, motivation, emotional intelligence, personality) and socio-technical aspects (e.g., mutual trust, adoption barriers, teamwork, collaboration, communication, coordination, project management). | |
| Responsible | Ethics-by-Design | Agentic SE needs to apply an ethics-by-design approach to its design and application, questioning the why, and not being driven by the why not. |
| Sustainability | Agentic SE needs to ensure sustainability of the core technologies, ecosystems, and the planet. | |
| Adaptive | Industrial Relevance | Agentic SE needs to adapt to new AI models and paradigms and work closely with the AI community and co-design with the industry. |
| Socio-technical Relevance | Agentic SE needs to predict and report socio-technical impacts of changes in AI models and paradigms. | |
| Foundational | Foundational Knowledge | Agentic SE needs to develop foundational knowledge about AAA across all areas, not only coding/programming. |
| Foundational Solutions | Agentic SE needs to develop foundational solutions to support AAA across all areas, not only coding/programming. | |
| Translational | Awareness and Education | Agentic SE needs to build awareness through education, certifications, funding bodies, media, and society; and maintain incident logs. |
| Actionable Guidelines | Agentic SE needs to study and enable the adoption of research knowledge and solutions into practice through usable guidelines and tools. |
Guidance for Agentic SE Vocabulary
The establishment of well-defined terminology is crucial for formalizing agentic SE. The paper suggests several criteria for designing and using vocabulary:
- Relevance: Terms must be fit for purpose and context within the agentic SE domain, with traditional SE vocabulary revisited for relevance.
- Coverage: Terminology should be conceptually complete and accurate, covering all goals, responsibilities, and artefacts of agentic SE process areas, beyond just coding.
- Acceptance: Terms must gain recognition and adoption by the community, ensuring humans agree to AI agents being given professional titles.
- Consistency: Uniformity of use across different technology and research groups, in both syntax and semantics, ensuring stability and adaptability over time.
- Philosophical alignment: Promote reflexivity, leading to awareness of underlying assumptions, power dynamics, and fairness, ensuring diverse voices are heard.
Adhering to these guidelines will help lay solid foundations for the emerging discipline of agentic SE.
Calculate Your Potential AI Impact
Estimate the transformative potential of advanced AI and agentic software engineering within your organization.
Potential Annual Savings
$0
Developer Hours Reclaimed Annually
0
Your Agentic SE Implementation Roadmap
A phased approach to integrating agentic AI into your software engineering processes, ensuring a smooth and effective transition across all lifecycle stages.
Phase 1: Vision & Ethical Alignment
Establish a clear 'whole of process' vision for agentic SE, including a robust ethics-by-design framework. Identify key stakeholders, potential impacts, and governance strategies.
Phase 2: Foundational Knowledge & Tools
Develop a deep understanding of human-AI collaboration patterns and socio-technical concerns. Implement foundational agentic AI tools and platforms, extending beyond just coding to areas like requirements and design.
Phase 3: Pilot & Iterative Adoption
Begin with pilot projects focusing on specific SE lifecycle areas (e.g., automated testing, design assistance). Gather empirical data, evaluate performance, and refine processes based on early feedback.
Phase 4: Scaled Integration & Training
Scale agentic AI solutions across multiple teams and projects. Provide comprehensive training for engineers on human-agent collaboration, new roles, and continuous learning. Ensure clear vocabulary and guidelines.
Phase 5: Continuous Evolution & Governance
Establish mechanisms for continuous monitoring of agentic AI system performance, ethical compliance, and socio-technical impacts. Adapt to new AI models and paradigms, fostering a culture of responsible innovation.
Ready to Transform Your Software Engineering?
Leverage our expertise to integrate agentic AI holistically across your software development lifecycle. Let's build a future where AI empowers human ingenuity.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat