Artificial intelligence is rewriting the rules of innovation, accelerating workflows, and opening possibilities once thought unattainable. But acceleration alone doesn’t guarantee success. Doug Sutcliffe, MIT Master of Science in Management of Technology and former Director of Data-Driven Marketing at L’Oréal, argues that the key lies in combining AI with Agile. Over the course of his career, including leadership roles at top corporations, his initiatives have earned major honors such as the Grand Prix Award for Data Application and the MIT $100K Entrepreneurship Competition. To a large extent, these milestones were achieved by flexibly adapting Agile to diverse contexts.
Now, Sutcliffe explains how Agile thinking can provide the discipline, structure, and adaptability needed to harness AI’s speed for lasting transformation, and to build AI products that move beyond novelty to deliver real impact.
Throughout my career, I have been able to lead numerous initiatives to award-winning results by grounding my work in Agile principles. For example, as a Product Lead for the Attribution Model on the precision marketing initiative that won the Grand Prize for Data Application, I developed a groundbreaking methodology that measured the impact of different marketing approaches and guided us toward the most effective strategy. The project established a proof of concept to leverage the brand’s digital assets, using retail partner data to analyze consumer purchase behavior, increase cross-sales, and deepen user knowledge.
My role was to apply an innovative and forward-thinking approach to data, setting new industry benchmarks by creating novel frameworks to compare marketing effectiveness across platforms. This strategy was foundational to the project’s success, which, in less than three months, proved its comprehensive power by increasing general purchase intent to 56% compared to just 18% for the control group.
Looking back, I see a common thread across such achievements. Their success did not come from simply applying a method or following a framework, it came from embracing Agile as a way of thinking and tailoring it to the unique context of each challenge.
This is more true than ever in today’s constantly evolving era of AI innovation, and I have been developing new principles to guide how Agile is applied in this world. Agile is about finding the right balance: delivering results in the short term while building the minimal critical structure to sustain long-term success. In a marketing analytics context, that meant balancing monthly and campaign-level deliverables with processes that would continue creating value well into the future. And on large, multi-faceted teams, it meant investing in technological transformation while ensuring every member understood not just what we were building, but why it mattered.
Why AI Raises the Stakes
With the rise of AI, this balance has become more critical than ever. AI accelerates everything. It makes it possible to automate workflows and scale experiments at unprecedented speed. But speed alone is not success.
Where Agile once focused on sizing projects into smaller, manageable components, AI shifts the challenge to sequencing. The real question is no longer “How can we break this into parts?” but “What should we do first, what should we leave for later, and what should we never attempt at all?” AI tempts us to do everything to generate every idea, automate every process, chase every opportunity. But doing everything means doing everything at the standard of AI, which in most cases, means doing it poorly.
Winning in this new landscape requires vigilant monitoring of AI’s evolving frontier and greater discipline than ever in constantly updating our heuristics of when to step away, when to harness AI’s support, and when to delegate fully.
Here, Agile provides a critical anchor. By applying Agile principles to AI development, we can create tools that are iterative, user-centered, and adaptable, which ensures adoption and impact rather than novelty for its own sake. At the same time, traditional core architecture investments remain essential to building an environment where AI can thrive. While these investments are now lighter and less costly, they still rest on the same principles I have always followed when balancing transformative “big bets” with steady, iterative evolution.
In turn, once these AI foundations are in place, they give Agile practitioners new ways to work faster, test more ideas, and eliminate low-value tasks. The result is a feedback loop where Agile accelerates AI, and AI supercharges Agile; an interconnected system of continuous improvement.
Case in Point: Turning Friction into Flow
One of the examples of this interplay is the Prompt Suite Chrome extension. The story behind its creation reflects a perspective I have always carried as a product owner: an obsession with the user experience and with removing every barrier that prevents people from adopting a better approach. With AI, I saw extraordinary potential paired with dramatic under-utilization. The same frictions that often stop customers from moving away from ingrained habits were also keeping people from unlocking AI’s full value.
Take prompting as an example. Entire articles have been written about how to engineer the “perfect prompt,” but that pursuit is really the equivalent of waterfall development — hours of upfront design that rarely survive first contact with the variability of real-world use. The real value comes from applying an Agile mindset: start with something workable, put it into practice, and refine it quickly through repeated use and iteration. The best prompt isn’t the most elegant on paper, it’s the one that evolves and improves every time it’s applied.
The problem with this, is that most people keep their prompts in Google Docs or OneNote lists, which are rarely opened. The extra clicks create just enough friction to discourage use. I wanted to eliminate that barrier, so I created Prompt Suite. By keeping prompts directly in the active browser window, available within any LLM, the tool makes it effortless to save, update, and reuse them. In doing so, it doesn’t just make prompts more accessible, it helps users build the Agile habit of continuous improvement.
Building AI Tools with Agile Principles
“AI accelerates, but AI + Agile transforms, turning faster coding into compounding improvement.”
The first version of Prompt Suite was built in just four hours with a single function: saving prompts for easy access in your browser window. That speed reflects what AI delivers on its own; acceleration. Writing code faster is like early Netflix offering quicker shipping on DVDs: an improvement, but still constrained by bottlenecks equivalent to AI’s brittle outputs and time-consuming manual regression testing.
Transformation comes when AI is paired with Agile. Instead of just producing more fragile code faster, I used AI to amplify continuous improvement within each sprint. The first feature I added was prompt import/export, which let me instantly reload prompts and folders after each update. That cut regression testing and iteration time in half, freeing me to ship more and better enhancements. Like Netflix’s leap to streaming, the same core technology yielded a fundamentally better development cycle; one that predictably compounded with every sprint.
In the same sense, Prompt Suite was not designed just to help users be more efficient by reusing prompts, but to help them adopt an Agile approach to AI in general. It encourages them to build the habit of continuous improvement by refining their inputs over time, while eliminating low-value tasks so they can devote more energy to strategy, creativity, and effective Agile management.
To me, this distinction is critical. Many tools let you schedule queries or watch AI perform tasks on your behalf. That’s like cruise control: it helps while you drive, but you’re still behind the wheel. Prompt Suite aims to be more like a Roomba: a tool that removes repetitive chores entirely, freeing you to focus on what only you can do.
In practice, people are using Prompt Suite for a wide range of tasks, such as:
- Meeting preparation: automatically scanning your calendar each morning, researching the people you’ll be meeting, and sending a well-cited brief before each appointment.
- Writing support: saving long blocks of context about writing style (for emails, reports, or LinkedIn articles) and reusing them across prompts.
- PR workflows: automating standardized media requests for large agencies.
- Developer efficiency: keeping debugging prompts, unit tests, and regex patterns in one library, ready to drop in anywhere.
- Student learning: streamlining study routines with ready-to-go “explain this concept” or “quiz me” prompts for any topic.
Scale, Store, Connect
However, the real opportunity lies in expanding these principles into a family of tools designed to bring what once required full enterprise software into lightweight, intuitive solutions for everyday users.
Here is an illustrative example. Traditional CRM managers often rely on complex platforms that centralize all contact information and workflow management. But non-professionals, whether job-seekers, freelancers, or small business owners, use these tools very differently. Their engagement is opportunistic rather than constant, meaning they don’t need every enterprise feature, but they do need the essential ones delivered in a way that fits naturally into their daily routines.
That philosophy has led me to begin work on three additional tools: PromptScale, PromptStore, and PromptConnect.
- PromptScale enables users to execute LLM prompts against long lists. Instead of manually running the same queries over and over, you can automate them across hundreds of items at once. That could mean enriching a large contact list, preparing personalized briefs for everyone you’ll meet at a conference, or running fifty different regression tests with one click after a software update.
- PromptCache makes it possible to save information from any screen without breaking your flow. Imagine you come across a LinkedIn post you know a contact would enjoy. With a single overlay button, you can capture that content, tag it “for Tom, from last week” and keep it stored until your next follow-up email. This provides a single reference point for relevant information encountered across chats, databases, and devices.
- PromptConnect brings these ideas together with targeted CRM features for non-professionals. It helps people stay mindful of the important relationships in their lives and makes it easier to nurture them over time. As I’ve seen, many job-seekers struggle not because of a lack of talent but because of missing connections and the employer trust required to transition into new roles. This is especially urgent in sectors like climate and clean energy, where the global skills gap has already surpassed 240 million jobs. PromptConnect is designed to help bridge that gap by enabling people to manage their networks more thoughtfully and consistently.
Such tools are all grounded in the same Agile leadership principles: deliver immediate value, remove barriers to adoption, and empower users to focus on what matters most. They are built on the conviction that AI will turn all of us into product owners, and that the tools we create must therefore be flexible, human-centered, and easy to use from day one.
A Path Towards Sustainability and Global Impact
The same principles extend to challenges of global scale. In clean energy, acceleration might mean installing more solar panels or deploying more batteries. But transformation comes from rethinking the system: redesigning grids for resilience, creating financing models that scale equitably, and building adaptive workforce pipelines that evolve with technology. Agile, when paired with AI, supplies the positive feedback loops that turn these complex transitions into cycles of learning and improvement, compounding over time instead of stalling out as pilot projects.
This systems-level perspective is one I deepened at MIT, where my research and thesis work focused on supply chain risks for fusion energy. MIT’s culture of systems thinking and its legacy of marrying academic rigor with entrepreneurial execution offer a powerful foundation for the next phase. Building on that foundation, I am now in the process of connecting with MIT and other leading universities to share these ideas more broadly, speaking with students, supporting their projects, and helping them apply Agile plus AI principles to maximize their impact in world-changing fields like clean energy and climate.
By linking entrepreneurial agility with academic rigor, we can move from accelerating experiments to transforming whole sectors. That means reducing waste across AI/ML development pipelines, enabling faster clean-energy transitions, and addressing the global climate and workforce crises in ways that are adaptive, scalable, and sustainable. AI can accelerate. Agile ensures transformation. Together, they can tackle the world’s most urgent challenges.
Still, it’s important to remember that as powerful as AI has become, it cannot replace strategy. Agile has always been about mindset. With AI, that mindset becomes even more critical. We must resist the temptation to do everything and instead focus on doing the right things, at the right time, with the right tools. Only then can we transform award-winning principles into global impact.
