(photo by Paul Kelly)
One of the great values of artificial intelligence (AI) technology is the ability to access vast amounts of data, perform analysis, evaluate alternatives, and make decisions. For project management this is a significant opportunity. Consider the implications for a change request where a new requirement is proposed and the project manager and project team consider the impact of the change on both the budget and schedule. They collect data and respond with the potential increase in cost and delay in schedule. In reality the change might affect numerous additional areas of the project.
In a situation of several change requests at the same time, the complexity of interdependencies may be too difficult to result in an accurate evaluation. An AI tool can manage both complexity and vast amounts of data and can identify a more accurate assessment of the impact of the changes on the project. Once a change is approved, the AI tool can update every project document to reflect the change.
However, developing and using any AI tool begins with data and that is the first significant issue. As data scientists begin to use big data (web link to definition) it is estimated that 60 to 80% of their time will be spent in data cleansing tasks. This includes detecting bad data, removing or correcting data and modifying data. If a business has subjected their critical business data to data standards and adhere to those standards then this provides a good start. With big datasets the lack of clean data means that most data fields will not move easily into an array for further analysis. The issues range from simple items such as having a data field with data in different formats, to more complicated items such as detecting valid or invalid data in a field. What is done with blank data fields? Mixed formats of items such as a date field with 'yy/mm/dd' or 'dd/mm/yy' are a nuisance but must nonetheless be addressed. Typos and improper capital letters make cross correlation of data another challenge.
Implementing AI and receiving the rewards is dependent on clean and manageable data. Is your organization ready to take advantage of AI tools? Probably not. Most organizations require a business case to get an ROI or payback on the cost and effort of data cleansing. That is likely to delay the benefits of AI technology and force the organization to seek incremental improvements instead of a game changing advantage.