Antonio NietoHBR Author | Director PMO | PMI Fellow & Past Chairman | Professor | Thinkers50 | Projects & CompanyBrussels, Belgium
AI often requires a different kind of data infrastructure and expertise. How can project managers balance the agility needed for AI experimentation against the stability required in traditional project management? Saving Changes...
Senior Projects Manager | Field & Marten AssociatesNew Westminster, British Columbia, Canada
Antonio, this is a very interesting question. Finding the equilibrium between agility for AI experimentation and the stability demanded by traditional project management can be a delicate balancing act. There are two important strategies that I can think of:
1) Deploy an adaptive approach which allows adjustments during the project lifecycle. This crucial for AI experimentation.
2) Divide the project into phases. Allocate specific phases for AI experimentation while ensuring stability in other phases dedicated to core functionalities.
...
1 reply by Kwiyuh Michael Wepngong
Dec 07, 2023 7:06 AM
Kwiyuh Michael Wepngong
...
Thanks, for this Rami
Saving Changes...
Luis BrancoCEO| Business Insight, Consultores de Gestão, LdªCarcavelos, Lisboa, Portugal
Dear Antonio This topic is very interesting.
Changing the paradigm is the solution. This paradigm shift reflects the need to abandon expired mental models and embrace dynamic solutions. Saving Changes...
Markus KopkoAI Enabler for Project & Program Mgmt | Founder PMotion.ai / The PM
AI Coach| PMotion.aiHamburg, Hamburg, Germany
Dear Antonio,
Balancing the agility needed for AI experimentation with the stability required in traditional project management is a nuanced challenge, especially given the unique requirements of AI in terms of data infrastructure and expertise. As a project manager, here’s how you could approach this balance:
1. Adopt a Hybrid Project Management Approach:
Agile and Traditional Blend: Use a hybrid model that combines agile methodologies (favorable for AI experimentation) with traditional project management approaches. This allows for flexibility and innovation while maintaining structure and control.
Iterative Processes: Implement iterative development cycles for AI components, allowing for experimentation and quick adjustments, while keeping the broader project under a more structured management approach.
2. Establish Clear Objectives and Boundaries:
Defined Scope for Experimentation: Clearly define the scope of AI experimentation within the project. Establish what is open for exploration and what needs to adhere to strict timelines and budgets.
Risk Management: Incorporate specific risk management strategies for AI-related aspects of the project, acknowledging the experimental and unpredictable nature of AI work.
3. Foster a Collaborative Environment:
Cross-Functional Teams: Build teams that include members with AI expertise and those with traditional project management skills. This fosters a collaborative environment where stability and agility can coexist.
Regular Communication: Ensure frequent and transparent communication among all team members to align on goals, progress, and challenges.
4. Invest in the Right Infrastructure:
Flexible Data Infrastructure: Invest in data infrastructure that is adaptable and scalable, allowing for AI experimentation without disrupting other project components.
Cloud-Based Solutions: Consider cloud-based services which offer scalability and flexibility for AI workloads.
5. Training and Skill Development:
Upskilling Teams: Provide training for project team members to understand the basics of AI, its implications, and how it integrates with traditional project methodologies.
Expert Consultation: If necessary, bring in AI experts for consulting and guidance.
6. Incremental Implementation and Testing:
Pilot Projects: Start with small-scale pilot projects or proofs of concept for AI initiatives before full-scale implementation.
Continuous Feedback Loop: Implement a continuous feedback loop, allowing learnings from AI experiments to be integrated into the project planning and execution process.
7. Monitoring and Adaptation:
Agile Monitoring Tools: Utilize project monitoring tools that support agile methodologies to keep track of the AI experimentation progress.
Adaptability in Plans: Be prepared to adapt project plans based on the outcomes and learnings from AI experiments.
Conclusion:
Balancing agility for AI with stability in project management requires a nuanced approach that embraces flexibility, clear communication, and a willingness to adapt. A hybrid project management approach, combined with the right team, infrastructure, and iterative processes, can enable project managers to effectively manage this balance. This approach not only accommodates the innovative nature of AI but also ensures that the overall project remains aligned with its objectives, timelines, and quality standards.
Financial Management Specialist | US Peace CorpsYaounde, Centre, Cameroon
Dec 06, 2023 7:37 PM
Replying to Rami Kaibni
...
Antonio, this is a very interesting question. Finding the equilibrium between agility for AI experimentation and the stability demanded by traditional project management can be a delicate balancing act. There are two important strategies that I can think of:
1) Deploy an adaptive approach which allows adjustments during the project lifecycle. This crucial for AI experimentation.
2) Divide the project into phases. Allocate specific phases for AI experimentation while ensuring stability in other phases dedicated to core functionalities.
You've answered your own question - as experimentation is a core component of any current AI initiatives, a traditional predictive approach won't be suitable for anything other than the implementation of a well defined, thoroughly researched and experimented AI use case.
An adaptive approach will be needed but not a random walk to nowhere - there needs to be a goal and set vision for each AI research initiative.
Experimenting with AI during a project is no different than with any other new technology. It is simply a tool used to execute the statement of work sometimes called enabling architecture. A few points to add to the excellent input above:
• Unless there is little new and novel in a project, they are never purely predictive. Even using the traditional Systems Engineering V model of top-down decomposition and bottoms-up verification, incremental changes occur at each step although the refinements become smaller.
• How you plan to use the tool will greatly affect where and how much you can experiment. If you are using AI to help draft communications, you can experiment throughout the project. If you are using it to develop a product architecture the cost of change increases rapidly, and your window of opportunity is much smaller.
I suggest performing some rough cost/benefit analysis to gain an understanding of how late into the project you can continue experimenting before the cost of implementing changes based off your findings exceed the benefit gained from the new capabilities. Saving Changes...
Community Champion
