A day using 4th Gen AR AI-ML in Construction - Use Case B

Megan visits the site using her new 4th Gen Augment Reality (AR) headset, passing through the first floor of a facility—a mall located outside Kissimmee, Florida. As she walks, blinking images appear, overlapping different elements.

As such, she sees existing rebars emerging from the slab, surrounded by an image of a wall colored in yellow. This AR-wall image represents the BIM element that was supposed to be completed last week. The blinking signal means that the element has more than 5 calendar days without leaping to the other completion status, remaining in  Work-In-Process (WIP) state. 

This is not the only WIP blinking AR element. On-site, Megan finds some constraints: materials like cement bags, fittings, and other ad-hoc storage, which make it impossible for the crew to complete the walls. She takes a photo using her headset, categorizing these materials as internal constraints since they belong to the contractor. At that moment, her deputy Joshua appears and asks for help in defining priorities for release inspections for columns a few yards ahead. Megan first inquires about the walls around her, and Joshua explains that even if the area were clear for erecting the walls, there wouldn’t be enough personnel available in the following 7 days to carry out the work. 

Megan adds this information to her AR pop-up form, which automatically returns the completion date updated for this section with a four-day delay. Her remarks also appear in the concerning BIM model, which also displays a heat map highlighting additional constrained elements, with the data shown in the headset as well. Here, the more constraints there are, the hotter an element appears.

The two then arrive at the columns that are about to be concrete casted. Megan, using her headset, updates the status to indicate that pre-casting preparation is complete and give the order to proceed. While the columns remain in a WIP state, the inspection point (completion node) changes its status in the completion database linked to the BIM model.

Megan is not the only one adding information; other supervisors, site engineers, subcontractors, and laborers are also updating the status of elements, either using headsets or directly through the BIM model. All these updates affect the overall status of deliverables, indicating a pattern of evolution that is analyzed by a Markov Chains statistical model to compute the completion date forecast in realtime.

The built-in headset software also incorporates shared worldwide BIM metadata for common elements, including constraints and rework statistics. It then compares patterns from Megan’s project to those from other projects and assesses the probability of delays and additional constraints. This Machine Learning information is displayed on the headset lens as well as on the BIM cloud-based dashboard. The statistics are shown using various benchmarks, such as global, regional, and national levels. 

Blockchain technology ensures a robust and shared BIM data community, which continuously feeds data into and draws information from the comprehensive BIM Analytics cloud to which the headset is linked.

Upon arriving at the office, Megan notices Solomon, the project control manager, waving to call her over to view the automated update of the main BIM model. The model displays the current status of project elements: completed elements are shown in green, work-in-progress (WIP) elements in yellow, and those not yet started in grey. The cloud-linked dashboard provides an analysis of the risk of falling into the '90% Syndrome,'  including a projected date for this occurrence, as the traditional progress curve approaches the upper completion limit (Q6). Megan also requested the customized dashboard Solomon had prepared to complement the official report. This customized dashboard highlighted the most probable WIP elements that could be completed soon, representing opportunities for quick wins in the pursuit of early project completion.

The AI assistance integrated with the headset software is synchronized with the project management databases, enabling the identification of root causes for various constraints, as well as potential and confirmed rework. These databases are also linked to the well-known Advanced Risk Management (ARM) system via a plug-in developed by ARBIM, which facilitates the synchronization and tracking of findings for further analysis within ARM.

The information fed through the headset allows the BIM model to highlight zones that are 'made-ready' (free of constraints), delivery-ready, high-potential delivery-ready, as well as zones with significant constraints. The number of constraints and the results from the constraint-matrix analysis enable the generation of stochastic indicators, which identify zones that exhibit results close to randomness. These results indicate either inadequate planning or poor execution.

This routine became familiar and productive, yet intense, especially in root cause analysis, as it demands validation rather than speculation. The hardest aspect of project challenges was changing the site engineers' mindset of considering bulk quantities as proof of progress and completion simultaneously, while progress does not necessarily equal completion, as completion means that work-in-progress elements equal zero.

During the monthly meeting, Isaiah, the project director, reviewed the latest project completion estimates, which were also included in the hourly automated flash report everyone received. Solomon raised some points about taking advantage of recently identified potential delivery-ready zones. Based on a simulation run, he suggests that if labor assigned to activities with time float were temporarily reassigned to these zones, it could significantly increase the chances of meeting completion targets earlier. According to the BIM ML model, this strategy could accelerate the completion date by five days.

The meeting concluded, and shortly after, Isaiah received a call from Jakob at ARBIM, the consultancy company that provided the headset. ‘Hey mate! The 4th Gen AR trial is about to expire… you ready to upgrade to the premium version of BIM site management?’ Jakob asked. Isaiah couldn’t be more convinced.