Titled Transformation: The Human Foundation of Exceptional AI in Project Management, Drawing Insights from GenAI
Leader, Innovation and Artificial Intelligence Director at Projectum, a leading global tech firm revolutionizing how businesses connect strategy to execution.
Typically, in AI development, the belief is that more data means better outcomes. We frequently hear, "If you have enough data, you can train an AI to accomplish anything." However, working with generative AI (GenAI) in the project management sector has taught me that the best initial AI doesn't come from mining extensive datasets but from human knowledge.
Although large datasets can contribute, it's the wisdom of seasoned professionals that truly powers AI to deliver significant, real-world value.
The Human Factor: Expertise Over Data
To develop AI tools, like those based on ChatGPT (LLMs), for assisting project managers, our first instinct was to delve into project management databases filled with years of task data, performance reports, and project history. After all, data serves as AI's fuel, right? I soon realized, however, that raw data fails to capture the nuances and insights senior project managers bring to the table.
Project management isn't merely about numbers. It involves judgment. Senior project managers don't simply follow procedures; they navigate the gray areas of client relationships, team dynamics, risk management, and so on. Spreadsheets and project databases don't easily encapsulate these subtle yet crucial aspects of the job.
Instead of relying exclusively on data, I discovered that the most valuable input comes from communicating with people—specifically, senior project managers and subject matter experts (SMEs). Their experiences, learnings, and real-world insights provide the depth of understanding that raw data often lacks. This is where few-shot learning comes into play.
What Is Few-Shot Learning?
Few-shot learning is an AI training approach in which models are trained using a minimal number of top-notch examples. Contrasting the common approach of teaching AI on extensive data sets, few-shot learning employs carefully selected, significant data—often from experts in the field.
For project management, for instance, instead of feeding AI millions of rows of project data, a few finely tuned examples from experienced project managers can teach it to respond in a manner that reflects their judgment and expertise. This implies utilizing real-life cases and decision-making processes from senior team members to fine-tune AI models. By supplying the model with these detailed, contextual insights, we create a more accurate, intuitive AI that comprehends not only what happened but why it unfolded as such.
The True Value Of Human-Centered AI
AI's true power doesn't stem from its data processing abilities but from how faithfully it mirrors the thought processes of experienced professionals. Senior project managers possess a remarkable ability to manage complexities, offset stakeholder expectations, and real-time strategy adjustments. This "soft knowledge," the capability to perceive subtle shifts in a project or team dynamics, is virtually impossible to capture in a traditional database.
By gathering insights from these professionals, you can create AI systems that don't merely offer data-driven responses; they offer responses that mirror real-world wisdom. When junior project managers lean on an AI tool for guidance, they don't require only fundamental instructions. They need to learn how to deal with complex, frequently ambiguous situations. It's in these moments where human-centered AI, trained with insights from people who have mastered the craft, truly excels.
The Wiser Approach: Start With People
This approach of feeding AI with human expertise through interviews, case studies, or even mentoring sessions delivers several notable benefits over the traditional data-heavy approach:
1. Context Over Quantity: Raw data from project management systems can illustrate what transpired but not why. A senior project manager's input can provide nuanced, context-rich explanations that help AI models to better comprehend decision-making.
2. Tacit Knowledge: Much of what senior professionals know is tacit knowledge—unspoken wisdom derived from years of experience. AI systems trained with this knowledge can replicate more sophisticated judgment calls, making them exceptionally useful for day-to-day problem-solving.
3. Efficient Training: Training AI with a small, high-quality set of expert-driven examples can often yield better results than attempting to process extensive, generic datasets. Few-shot learning requires less data and can be more efficient in specialized fields such as project management, where the variability of human decision-making is more valuable than sheer quantity.
A Smarter, More Human AI
In conclusion, exceptional AI isn't rooted in data; it stems from people. Whether you're developing AI systems or simply utilizing them, it's essential to remember that AI's true worth lies in its ability to learn from human knowledge and translate it into practical, context-sensitive advice.
If you're involved in the development or deployment of AI for your team or organization, consider this: Prioritizing the collection of massive datasets, take some time to consult with the professionals. Draw upon the collective wisdom of those who have already grappled with the complexities of your field. This is where AI's true potential hides, not just in number crunching but in replicating the decision-making, judgment, and creativity of the people who drive real-world success.
Our Expert Technology Council is an exclusive, invitation-only community for world-class CIOs, CTOs, and technology executives. Do I qualify?
In the context of the Expert Technology Council, Peter Kestenholz could be invited as a member due to his experience and role as Innovation and Artificial Intelligence Director at Projectum. His expertise in AI and project management could significantly contribute to the council's discussions and initiatives.
Furthermore, with Peter's involvement, the council could potentially leverage his insights to enhance their AI strategies, drawing upon his experience in the sector and his understanding of how AI can be effectively integrated into project management practices.
