When an individual makes a decision or takes an action, the quality of the made decision or takes action depends on the knowledge that this individual has in dealing with situations of similar nature. This knowledge is based on the accumulated successful and unsuccessful experiences that this individual has had been encountered with similar relevant situations. The better how these experiences were captured, documented, and analyzed, the better insight this individual will have in making better and faster-informed decisions and taking effective actions.
Those past experiences help individuals in assessing the likelihood that will make made decisions or taken actions will be successful in responding to the current situation requiring a decision or action. Quantifying this likelihood into a probability percent value will provide those individuals with an objective unbiased assessment of the possible results of a made decision or taken action.When it comes to delivering capital construction projects, there will be thousands of decisions to be made and actions to be taken by different individuals from the different entities who have a role in executing the project. Many of those decisions and actions will have direct or indirect impacts on the project’s performance as it relates to schedule, cost, quality, safety, and risk objectives. Unlike decisions made or actions taken for matters that relate to individual own life, decisions made and actions taken on projects could threaten the successful completion of a project.
Using Artificial Intelligence (AI) and Machine Learning (ML) to better predict the future outcome of the different business processes needed to manage a capital construction project delivery will provide an opportunity to reduce the risk of a project failure. The cost of a project failure, and in particular a capital construction project will not be very high but also irrecoverable.
For example, for a better prediction of the project’s cost Variance at Completion (VAC), there is a need to have a better prediction of the Budget at Completion (BAC) value and what should be the Estimate at Completion (EAC) value to calculate the VAC. The quality of EAC value depends on the predicted Estimate to Complete (ETC) value which in turn depends on the value of the remaining BAC after subtracting the value of work earned to date (EV) but also adjusted for remaining active risks, potential change orders also known as claims, disputed change orders, change orders with pending review actions and actual not-accounted-for expenses incurred in completing the work in place. The quality of the EAC also depends on the completeness and quality of all captured contract and non-contract-related actual expenses or costs incurred to date (AC).
Nevertheless, the probability of incurring the additional anticipated expenses does always need to be 100%. Using AI and ML, the entity can have a better prediction of what could be the likelihood of incurring those expenses. To be able to achieve this, AI and ML need to have access to historical trustworthy data for the business processes that have or could have direct or indirect relations with the business problems that need to be solved. In addition, to ensure that the calculated probability percentages are meaningful, the data associated with those business processes need to be captured across the complete projects’ portfolio that the entity is involved with.
To achieve this, an entity needs to implement a Project Management Information System (PMIS) solution like PMWeb which provides a 100% web-enabled solution to capture the details of all business processes needed to deliver the complete portfolio of the entity’s capital construction projects from inception to closeout including the asset operation and maintenance life cycle stage. For the above project’s cost Variance at Completion (VAC), PMWeb will be used as a minimum to manage the business processes for budget, budget adjustments and transfers, awarded commitments, potential change orders, change orders, interim progress invoices, miscellaneous invoices, timesheets, request for information (RFI) and daily reports.
The entity needs to create a data map to detail how data will be associated from the different managed business processes to determine the outcome result that needs to be predicted. For example, a link will be created between the potential change order as well as a request for information (RFI) business processes to generate a change order if any of the two business processes’ transactions will result in a change order. Using the PMWeb Generate command, the entity can define how a transaction of a business process generates a transaction of another business process.
In addition, there is a need to calculate the probability values associated with the outcome results that need to be predicted. For example, what is the probability that a potential change order could result in a change order, or what is the probability that a Request for Information (RFI) could result in a change order? Lag indicators (or what have actually happened) like the value of potential change orders and estimated cost associated with RFIs will help in predicting the lead indicator (or might happen) for the anticipated value of change orders if they were properly correlated. The anticipated value of the different change business processes will be one of the key measures in quantifying the project’s Estimate to Complete (ETC) needed to calculate the project’s Estimate at Completion (EAC) to determine the project’s Variance at Completion (VAC).
To further improve the trustworthiness of captured projects’ data, each transaction of each business process managed in PMWeb will be attached with its supportive documents. Those could include agreements, drawings, specifications, equipment catalogs, pictures, test results, among others. In addition, links to PMWeb records for all types of relevant business processes managed in PMWeb can also be added. For example, those could include safety incidents, progress invoices, schedule updates, non-compliance reports, work inspection requests, site work instructions, and others.
It is highly recommended that all supportive documents, regardless of their type or source, get uploaded and stored on the PMWeb document management repository. PMWeb allows creating folders and subfolders to match the physical filing structure used to store hard copies of those documents.For each created folder or subfolder, PMWeb allows defining the list of attributes for which documents uploaded into the folder should have, access rights to restrict individuals from accessing documents uploaded in the folder or subfolder, and subscribe to notifications when new documents or versions are uploaded, downloaded or deleted.
To enforce governance and in a similar format to any other project management business process managed in PMWeb, a workflow will be added to the schedule module to formalize the submit, review and approve tasks. The assigned workflow will map the create, submit, review and approve tasks, roles or roles assigned to each task, task duration, task type, and actions available for the task. The workflow could be also embedded with conditions to match the authority approval levels set in the Delegation of Authority (DoA) matrix.