AI-centric Megaproject Delivery
Traditional construction project delivery performance suffers from long and well publicized negative global productivity lags: negative compared to the industry 20 years ago, and negative compared to other industries. This negative productivity is associated with both budget shortfalls and turnaround delays. Infamous megaprojects suffer from these same shortcomings to exponential degrees.
Megaproject delivery is a messy process that lacks standardization, checks and balances, proper oversight, and accountability. The sheer magnitude of documentation, redundancies, and fragmentation of information makes unruly megaprojects appear impervious to deconstruction. It does not help that parties sequester information to themselves, selfishly begrudging their building partners to promote their own agenda. Change is in the wind, and that change is AI-centric megaproject delivery.
Why the Industry Sucks
I have posted recently here about the 98% over-budget rate of megaprojects. One of the problems with ameliorating megaproject delivery is with the limiting linear structure, and the behavior of the primary entities:
Traditionally, the owner controls the of project cash flow, and makes decisions to facilitate his designers to generate constructible drawings. This is idealistic, if past performance is any consideration. Part of the problem is that designers are seldom scrutinized by ownership as builders are, and have no other authority to answer to. If the owner doesn’t know how to drive the designer, the deliverables will be late. If the Owner doesn’t make timely decision and progress payments the project may also languish.
“On a good day, better CMs will manage hapless designers, as well.
In my model (AI-centric) the CM and architect are on equal footing, and must provide transparency to all partners. Project deliverables – rarely mentioned in project schedules – would be properly registered and tracked, as would job change orders.
Since megaprojects can only help but disrupt communities in ways for the (mostly) worse, the public has a right to know what’s going on, and be involved. Megaprojects can utilize independent private sector NGO oversight consultants for estimating and scheduling, or project controls. These would be ownership schedules: CM will build an independent schedule from his subcontractor input.
P3 consultants are intended to enhance the level of project analytics and forensics without bias. They might include a consortium of representatives from the numerous groups with an interest, across various entities in the public and private sector.
Owner & Stakeholders Drag-down
With Ownership in between the builder and designer, you always have fertile ground for major conflicts, disconnects, and impasses that eat away at the integrity of the project delivery. The lines between authority and accountability become blurred, as the three play tag-team wrestling with the public and owner’s time and money.
Planning to Fail? A Typical Debacle
An owner contracts with a designer and with a construction manager for a fee based megaproject. The budget and timeline are aggressive, and would scare most GCs, but might be more suitable as a CM project delivery.
Accordingly, the CM contract is let at an aggressive rate, which the CM intends to later offset by passing along the aggressive buyer positions toward vendors and specialty trades. Going into the project, the subs already feel squeezed. Call it ‘poisoning the well,’ an unhappy impecunious marriage for everyone.
Once the project is shovel ready, the underendowed entities begin their work off on a bad foot: a paucity of design documentation and scarcity of money on the job. At the end of the day, an owner has created a contentious scenario to build a project. This phenomenon is painfully evident even to passersby: a building sorely undermanned has the air of an abandoned hospital or mausoleum, where little or no substantive activity is undertaken.
PMI to the Rescue
To its credit, the Trump administration passed the PMIAA intended to improve federal program management and project delivery techniques. This step in recognition of the Fed’s utter ineptitude at program management. Despite the anointing the PMI as sole taskmaster and drafter of the new protocols, the expected impact on the industry can’t be measured for some years after its enactment. Moreover, it only applies to Federal work
That means a lot of PMI consultants will be hard at work building theories and testing them out in laboratories. This will take time – to populate their database with a data set they can analyze and make their recommendations for improvements. It could be a long and tedious process. Facilitating the program once the recommendations are made will require government resources in the same way Obama Care did at the state levels, with the Fed providing the infrastructure to unwilling states governments.
In the meantime, AI shows potential as a robust and dynamic project co-delivery tool – one that could take over large components of program management processes that are traditionally delegated to hapless data-gather-and-entry clerks. In lieu of the traditional trickle down pyramid structure above, AI can take the place of facilitator where others have historically failed miserably. You’re starting to wonder if this is science fiction. It is not.
AI: the Moral Imperative
Having no conflict of interest, AI can serve as an impartial, indiscriminate, and equal opportunity interface for the project team, from the primary hub, down the line. AI doesn’t have any bedside manner and is not afraid to step on anyone’s shoes. All are treated equally under AI’s guidance, and there is nowhere to hide. Checks and balances + independent oversight makes for wholesome program management.
With AI calling the shots, there will be no place to hide ineptitude, egregious errors in judgment, or cases of graft and corruption, all of which have besieged the industry since before most of us can remember.
This democratization of the megaproject project delivery process will be a bitter pill for anyone who has made their living off its excesses and failings, however, the status quo is no longer acceptable.
AI-centric Megaproject Delivery
Primary Hub Front-end
AI-centric Megaproject Delivery is a hub-based system designed to equally facilitate four distinct primary entities, as shown at left:
Architect and Engineer
These entities will have equal access to all primary hub documentation. CM will facilitate secondary and tertiary project team members through the AI server. This arrangement dissolves disparities in degrees of control and accountability that traditionally favor owners and shield the AE team – leveling the playing field.
AI Semi- and Fully Automated Activities
AI-centric project delivery calls for a complete revamping of traditional project structures from trickle down hierarchy, to equal partnerships. AI will be able to facilitate semi-automated and fully automated tasks, as follows:
Manual, Semi-Automated AI, Automated AI
The above project documentation can be either semi- or fully-automated using AI-centric project delivery tools, effectively redelegating general requirement activities to automation. Such an adjustment or refinement will substantially reduce the cost of project delivery, increase levels of exactitude, create an interface or platform that treats all entities equally, without bias.
Construction Manager’s Role
A Construction Manager is a facilitator from design, through preconstruction, build-out, and close-out. He receives all documentation from resources and vendors, and uploads to AI server. AI logs, distributes, verifies receipt, notifies, and reminds all tagged parties. In other words, the only manual part of a CMs process is uploading and distributing documentation.
This might inspire many to worry about clerical and office personnel jobs. Just like schedulers and estimators, AI will always rely on operator interface and verification processes. Once AI starts doing all the redundant administrative work, the personnel who used to dedicate their time to mundane things will be liberated to take on more challenging and rewarding responsibilities.
Architect and Engineer Roles
The expectation in years to come is that AI will also take over much of the redundant and time consuming design processes through generative design. Beginning with preconstruction, Architects and Engineers will upload all project deliverables – BIM and CADD files, to AI for posting. AI in turn, will process the files for instant access to all Primary Hub entities. AI will correlate a schedule activity with all deliverables and mark progress semi-autonomously.
Architects and Engineers communicate with ownership and CM through the AI server, and let it worry about notification and distribution. This nuance transfers vast paper workloads away from personnel to free them up for more interesting and constructive activities.
Secondary Hub Front End
At the next node level down from the CM are entities the CM will facilitate and integrate with through the AI servers:
Resources will post to CM, who in turn posts to AI server.
Department of Buildings & OSHA
DOB and OSHA issues can be file managed by the CM and through the AI server. These organizations are already digitized.
CM is responsible for managing all finances on the project through the AI server. This includes job accounting and progress reports for bank draw-down.
Brave New Megaproject
By relinquishing executive control and assigning it to an impartial facilitator – AI, we can eliminate mega-waste and unnecessary mega- delays, in the issuance of design documentation, payments, RFI responses, and other job critical instruments that traditionally impede progress. This transparent scenario is in stark contrast to Clueless Emperor trickle down decision making.
With AI as Paymaster, Owners can no longer delay and dilute resource progress payment applications at their own whims and caprices. AI would review and validate progress payment apps, and guarantee that resources are paid in a timely manner, not when the owner feels like it.
With AI as Clerk of the Works, it will be virtually impossible to lose track of current and previous drawing issues. All primary and secondary hub members are automatically notified of every drawing, sketch, and RFI response. AI creates the drawing logs, places reprography orders, prepares transmittals, and issues the drawings to all parties.
With AI as Clerk of the Works, designers will be required to make timely project deliverables as plotted and tracked in the master schedule, where they traditionally have been no-shows. With AI tracking and notification there will be no place to hide, or while away heinous volumes of billable hours.
It’s the general consensus that artificial intelligence will permeate many industries, and already has in more advanced countries. The ROI of AI-centric project delivery is too good to pass up, and will only become better, as the technology develops and becomes more mainstream. The failure of megaprojects to deliver is the impetus. Such great responsibilities can no longer be trusted to a few elite plutocrats or erudite officials.
However distant most of AI’s impact – it will be felt in increments, as AI becomes more sophisticated. What I am saying today is that most of what is needed to better manage megaproject delivery is closer at hand than we may think. Why not let AI handle the load of redundant digital transactions, thereby eliminating vast waste at the administrative and executive level.