Clarity

 View Only

Radio Clarity: AI Hype vs. Reality in PPM and SPM

By Brian Nathanson posted 24 days ago

  

The buzz around AI is undeniable. ChatGPT captured the world's imagination a couple of years ago now, making people believe AI is ready to revolutionize everything, including PPM and SPM, but as I recently discussed with my colleague Alf and Jason, the reality is more nuanced.

While AI has made significant strides, particularly in language manipulation, it still faces significant limitations that prevent it from being a "magic bullet" for our industry.

Here's a breakdown of our conversation and what it means for Clarity:

The AI Strengths We Can Leverage:

  • Narrative Generation: AI is fantastic at generating narratives, summaries, and improving existing text. We're already doing this in Clarity with features like text suggestion and roadmap summarization. We can build upon this by enabling AI-powered document generation for creating business cases and streamlining status reports. Imagine, for instance, using AI to automatically create a compelling narrative summarizing a month's worth of status reports for a monthly deck, rather than manually consolidating information.

The AI Challenges We Need to Address:

  • Limited Understanding of Numbers and Dates: Large Language Model (LLM)-based AI struggles with anything involving numbers, dates, or sequences. This means it's not yet ready to fully handle budgeting, resource allocation, or forecasting based on time-series data. For example, AI can't yet understand the significance of having spent $600,000 in the first month of a year-long $1,000,000 project versus having spent the same amount two weeks before completion.

  • The Memory Problem: AI agents in general suffer from "short memory" and struggle to retain context across multiple interactions. This hampers their ability to learn from past experiences and provide nuanced advice. Imagine a virtual project manager AI trained on the PMBOK guide. It can access this guidance, but without a durable memory, it might struggle to retain both its training and the entire project history, leading to less informed recommendations or, worse, complete nonsense because it can only remember all the details about the last 6 weeks of a 1 year effort.

  • The "God Button" Myth: While there's a desire for an AI-powered "God button" to automate everything, this is unrealistic given the current limitations. We need to be cautious about overselling AI capabilities and focus on delivering practical solutions. For example, while customers might dream of an AI that automatically allocates resources and predicts project success, the current lack of reliable data, trust in AI outputs, and general unwillingness to both send and receive data from external sources makes this impractical for many organizations.

Looking Ahead: The Path Forward for Clarity

  • Global Search is Key: We're focusing on building a powerful global search engine, which will be foundational for any AI, as it needs to understand where our data lives to be effective. Imagine an AI agent that can access all your Clarity data, not just fields on a per-object basis. This will allow it to provide more comprehensive insights and recommendations.

  • AI Agents for Specific Tasks: We're aiming to develop AI agents targeted at handling specific tasks within Clarity, like providing coaching on best practices and suggesting missing activities - without actually attempting to replace the role entirely. For example, an AI agent could suggest key steps for a new project manager to get a project off the ground, based on the organization's specific processes and best practices, without actually trying to usurp the project manager's job.

  • Addressing the Context Challenge: Solving the context problem is paramount for AI to truly mature. This involves building AI agents that can retain information across multiple interactions as well as understand the larger organizational context. This means developing AI that can learn from past projects, consider organizational priorities, and even take into account external factors like industry news or economic trends.

The Bottom Line

AI holds tremendous potential for PPM and SPM, but we need to be realistic about its current limitations. By focusing on practical solutions and leveraging AI's strengths while addressing its weaknesses, we can deliver real value to our respective customers and be viewed as leaders in the evolving AI landscape.

0 comments
7 views

Permalink