PUBLISHED DATE: 2025-01-13 04:04:50

The AI-Assisted Agile Manifesto | Publicis Sapient

Evolving Agile for the AI Era: Introducing the AI-Assisted Agile Manifesto

Arguably the most universal language among software developers, even surpassing programming languages, is the agile methodology, rooted in the principles of the Agile Manifesto. The Manifesto for Agile Software Development, originally crafted to improve software development through collaboration, prioritizes flexibility, teamwork and adaptability over rigid processes. These guiding principles have since extended beyond software into fields like design and user experience, sometimes morphing into diluted or altered versions that stray from the manifesto's original intent.

As artificial intelligence (AI) becomes deeply woven into the fabric of software engineering, it’s not enough to create yet another version of agile with new buzzwords. The core principles of the Agile Manifesto must be genuinely updated to reflect the evolving landscape. The manifesto was initially built around human collaboration. At Publicis Sapient, we recognize that the future of development requires us to collaborate not only with people but also with AI agents, tools and platforms. Success now hinges on how well we treat AI as a vital partner—not just an optional tool but a first-class teammate.

To meet this challenge, we propose an evolution: the Manifesto for AI-Assisted Agile. Our goal is to offer a framework that guides agile practices into the AI-powered future—a set of principles designed to keep pace with the rapidly changing world of software development—much like the founders intended.

Author
Rakesh Ravuri
CTO, SVP Engineering

What is the Manifesto for AI-Assisted Agile?

The Manifesto for AI-Assisted Agile is an evolution of the Agile Manifesto for software development assisted by AI. The original Agile Manifesto, created by a group of software developers, is as follows:

“We are uncovering better ways of developing software by doing it and helping others do it. Through this work we have come to value: That is, while there is value in the items on the right, we value the items on the left more.”

We propose the Manifesto for AI-Assisted Agile, as follows:

“We are uncovering better ways of developing software by doing it with the assistance of generative AI tools. Through this work we have come to value: That is, while there is value in the items on the right, we value the items on the left more.”

Explaining the new Manifesto for AI-Assisted Agile

We carefully evaluated each principle to determine if, why and how it should be changed to reflect collaboration with AI agents, tools and platforms in the software development lifecycle (SDLC).
  1. Individuals and interactions over processes and tools
    Individuals and AI interactions over rigid roles and ceremonies
    The original manifesto’s ethos of people over process does not change in the AI-assisted manifesto, but we believe that AI interactions are just as essential as human interactions to improve these aspects of software development:
    • Flexible roles and adaptation
      AI companion tools and training allow for more fluid roles and more skills per individual
    • Cross-functional collaboration
      AI-driven efficiencies allow more time for creativity and innovation between functions
    • Enhanced decision-making
      AI can make unbiased decisions, using insights to provide maximum value
    • Adapting ceremonies and processes
      AI can run ceremonies
    • Reducing admin
      AI can drive communications and reporting
    "Individual’, interactions with AI personas in the team will become more important... if we are not considering AI as a full-fledged, first-class citizen of the process, you are missing the point."
    Rakesh Ravuri , CTO at Publicis Sapient
  2. Working software over comprehensive documentation
    Explainable, working software over comprehensive documentation
    If AI makes code explainable and auditable, the need for comprehensive documentation is greatly reduced. In the future, most code will be AI-generated rather than handwritten, necessitating an explanation for how it works. This shift moves us from traditional documentation, which shows how software functions, to explainability, which reveals why the software behaves in a certain way. AI achieves this through several mechanisms:
    • AI-written and governed code is inherently auditable and accurate, ensuring that every line of code is transparent and traceable
    • Segregating code helps keep code organized to be regenerated, while AI-assisted prototypes allow for quick iterations and improvements
    • AI-driven demos and training sessions provide practical insights into how the code operates
    Additionally, AI can generate real-time explanations of live code, giving developers an immediate understanding of how systems are functioning. By collecting real data, AI offers rapid feedback, allowing for adjustments that enhance performance and reliability. This approach bridges the gap between human and AI-generated code, making the software development process more transparent and efficient. AI assistance leads to quicker development cycles, enabling the early release of valuable features and reducing the time to market.
  3. Customer collaboration over contract negotiation
    Valuable solutions over contract negotiation
    The original manifesto’s ethos of collaboration remains the same, but AI lets us bring value directly to the discussion. In the past, we’ve focused on fulfilling customer requests through collaboration. However, AI allows us to efficiently validate and ensure that these requests genuinely add value to customers. As the saying goes, if you ask customers what they want, they might just ask for a faster horse and cart—they don’t always know what’s possible with AI-powered technologies. By implementing A/B testing and generating insights, AI can help prioritize the backlog based on data-driven decisions rather than customer opinions. AI excels at uncovering valuable insights that humans may overlook due to biases and preconceived notions, enabling a more effective and objective approach to prioritization. We believe AI will help clients come to value not just our teams but our tools as well.
  4. Responding to change over following a plan
    Responding at pace over perpetuating legacy patterns
    The core facet of this value is to predict change, at pace, over solely reacting to the present. Merely responding is no longer enough—responding at pace is the new standard of excellence. Consider this example: When a customer tweeted about an issue, the company's leadership responded within six hours, addressed the problem within six days and implemented a solution within six weeks. It’s not just about fixing bugs; it’s about how quickly they are fixed. To keep pace, it’s essential to build systems that enable faster responses, such as automating processes to accelerate workflows. Traditionally, a new feature would be added to a story, picked up by another team, developed and then deployed. However, using generative AI to automate steps—like auto-updating stories, regenerating, deploying changes and even ways of working and ceremonies—can drastically reduce human intervention and save time.

Sapient AI for Applications

Explore integrated AI solutions that combine cutting-edge technology with human expertise to deliver unmatched software development efficiency and quality.
Learn more

How we are transitioning our organization to the Manifesto for AI-Assisted Agile

While we are still in the early stages of AI-assisted software development, we are already figuring out ways to integrate the principles of the Manifesto for AI-Assisted Agile into our organization.

Upskilling

Actively upgrading the skills of our team members through targeted training programs, hackathons and workshops. These initiatives focus on how to effectively leverage generative AI systems like GPT models. We are teaching our teams how to fine-tune these models, as well as introducing them to the basics of data engineering and data science. This skill upgrade is crucial for enabling our workforce to maximize the potential of AI technologies.

Effective tool usage

Additionally, we are emphasizing the importance of effective tool usage. Through specialized training, our team members are learning to utilize our proprietary Slingshot tool and other AI-driven platforms like GitHub Copilot within the SDLC. These training sessions cover a wide range of generative AI-based tools that enhance DevOps processes, equipping our team to use these technologies more efficiently and effectively. Tools are becoming essential in this new approach, and our goal is to ensure our workforce is adept at using them to their fullest potential.

Behavior and mindset changes

Beyond skills and tools, we are focusing on the behavioral dimension of this transition. Adopting AI-assisted agile methodologies requires a significant mindset shift, particularly when embracing the "Triple A" approach—AI-augmented, AI-automated and AI-assisted. For example, using tools like Copilot encourages developers to adopt a polyglot mindset, as they can prompt AI to generate code in multiple programming languages. This shift encourages creativity and challenges past assumptions about what is possible. If asked to create something as sophisticated as a Google-like search box, our teams now have the confidence to explore solutions that were previously deemed impossible, blending traditional algorithms with AI capabilities to solve problems more creatively.

The future of the Manifesto for AI-Assisted Agile

The software community is encouraged to adopt and refine practices and behaviors related to the Manifesto for AI-Assisted Agile to ensure that agile practices remain relevant and effective in an AI-driven world. Embracing these principles will allow agile methodologies to evolve alongside technological advancements, maintaining their impact and usefulness in software development. With the integration of AI, we are entering an era where the way software is developed will fundamentally change. AI technologies will enable us to create software faster, with greater variety and at a lower cost. More people will have the capability to develop software, and those with advanced skills will be able to achieve even more ambitious outcomes. This shift promises better, faster and higher-quality software, opening new possibilities for innovation and creativity in the field.

The Manifesto for AI-Assisted Agile outlines the principles that will guide the evolution of these processes. It encourages us to evaluate new tools and approaches by asking key questions, such as: By adopting these principles, we can ensure that agile practices remain adaptive, value driven and aligned with the future of AI in software development.

Rakesh Ravuri
CTO, SVP Engineering
Let's connect