Improving project management with AI and analytics at KMWorld Connect 2020
At KMWorld Connect 2020, Art Murray, CEO the consultancy Applied Knowledge Sciences, Inc., reflected on the need to combine human and machine intelligence in ways that result in more efficient and effective problem identification, prevention, and resolution.
Murray's presentation, titled "How Did We Miss That? Detecting and Responding to Weak Signals at the Working Level," was part of the AI in action session, in which Allison Bushell, director, analytics, and Fang Chen, principal analytics consultant, Confirmit, also presented.
Common problems with complex project management, said Murray, are that they are
- Rear-view mirror-oriented (not forward-looking)
- Alerts often generated far downstream from root causes/conditions
- Key insights regarding project status at the activity level are often lost at the executive decision-making level
- Complexity of the project organization, tasks, and workflow contributes to confusion, miscommunication, and error
- Traditional Earned Value Management (EVM) system disconnected from systems engineering, resulting in little or no insight into failure modes
As part of his talk on project intelligence, Murray used transcripts made publicly available by Boeing about its 737 MAX following two fatal crashes and subsequent grounding (it has since been approved by the FAA to fly again).
Using the discourse surrounding the aircraft for a text and sentiment analysis study, Murray considered what analytics software was able to identify in the project discussions versus what three experts uncovered. The goal was to identify lessons that may be applicable to other situations in order to generate recommendations and solutions for the use of text analytics.
Murray's slides from the presentation are available now on the KMWorld Connect 2020 website and a replay of his talk will be made available in the next few days.
Traditional project management
According to Murray, traditional project management tends to be rearview mirror-oriented. Large volumes of data are often filtered, summarized, and time-lagged, with little insight into the many small events at the activity level that collectively impact overall performance.
Murray walked attendees through elements of the machine-aided text analysis, discussing where the red flags were.
The ability to time-tag comments is important in analysis, he noted. If you detect signals early enough, you can make changes faster and save time and money, said Murray.
In projects, correcting errors and re-work can reach up to 80% or more of the total lifecycle cost of a system. By combining human and machine intelligence, organizations can achieve more efficient and effective problem identification, prevention, and resolution.
Just as important as being able to pick up cues from words in text is being able to identify a lack of specific words being mentioned, which can also in itself be a red flag. While machines make it possible to go through much more text than humans would be able to cover, machines can also miss sarcasm and external references that humans would understand and pick up on. Humans are also more adept at understanding misalignment of team objectives and failure to take action.
Automation with people in the loop
In conclusion, said Murray:
Semantically-rich text analytics is difficult to achieve
- You can’t put 3-5 decades of experience-based intuition into an ontology overnight, but
- You can almost always add something to the machine knowledge base every time you use it
Current tools still don’t scale well computationally for large volumes of unstructured text
- Possess intuition, insight, and foresight
- Understand basic human traits (biases, emotions, etc.)
- Purely computational
- Don’t understand meaning, purpose, intent, etc.
- Can’t formulate evocative questions
Ultimately, Murray said, what is needed are automated text analytics and also human analysis for sense-making, as well as governance and oversight, to manage processes.
It takes strong human intervention to align text topic categories to your business decision areas