
The limits of traditional risk reviews in US transit programs
How major transit organizations are rethinking delivery risk and the new urgency behind it.
New scale, old approaches
Over the past decade, US transit agencies have embarked on some of the largest capital investment programs in their history. Projects like LA Metro’s D Line Extension, New York’s Second Avenue Subway Phase 2, and the Gateway Tunnel are multibillion-dollar undertakings with intense political visibility and public scrutiny.
The challenge? While the size and complexity of these projects has exploded, risk management practices in many agencies have barely changed: Manual schedule reviews, static spreadsheets, and inconsistent assessment methods are still the default, even on billion-dollar megaprojects.
This gap between project scale and risk management capability can lead to blind spots, late discoveries of critical risks, and costly mid-course corrections.
Where Manual Reviews Break Down
The most consequential decisions often happen early in the lifecycle in major U.S. transportation programs:
- Reviewing and approving baseline schedules
- Selecting contractors
- Making go/no-go calls on key phases
Yet, across many DOTs and transit authorities, these decisions still rely on:
- CPM reviews that take weeks to complete
- Risk registers locked in static spreadsheets
- Methods that vary across bidders, contracts, or delivery teams
The result: On paper, these processes tick compliance boxes, but in practice, they leave leaders making billion-dollar decisions on outdated, assumption-heavy inputs and expose project teams to risks that could be seen and mitigated earlier.
What We’re Hearing from Major Transportation Authorities and DOTs
In recent engagements with U.S. transportation authorities, including a major West Coast agency delivering large-scale light rail expansions, we’ve heard a clear pattern:
- “Our capital program has outgrown our current risk review process.”
- “We need more transparency and defensibility in how we evaluate contractor bids”
- “It takes too long to surface risks we should have seen earlier.”
These teams are actively exploring ways to modernize how they assess and forecast delivery risk, particularly on politically sensitive projects involving Progressive Design Build, multiple bidders, and complex interfaces.
A Shift in Risk Management Approach
Forward-looking agencies are beginning to adopt a new standard for delivery assurance, leveraging AI to change the pace and precision of how risks are identified, communicated, and mitigated.
Where manual reviews can take weeks and still leave blind spots, AI can analyze an entire schedule in minutes, scanning every construction activity, surfacing hidden risks, and highlighting potential issues buried in complex dependencies or spread across multiple projects. This analysis can be repeated as often as needed - weekly, monthly, or at key milestones - without pulling scarce human resources away from other priorities.
The result is a new division of labor where machines handle the heavy computational work of risk analysis, while human experts focus on strategy, decision-making, and negotiation. The outcome is both broader and more in-depth insight than traditional methods can offer.
In practice, AI-enabled risk reviews mean:
- Faster, repeatable analysis that supports real-time decision-making
- Greater transparency and defensibility in bid evaluations
- The ability to manage risk across entire portfolios, not just projects
As one senior program leader put it:
“It’s the missing link between our risk workshops and the real decisions that drive project success.”
Why Transportation Program Leaders Should Care
The political and public environment for U.S. transit agencies is unforgiving. Large-scale delays or overruns can erode public trust, trigger oversight investigations, and jeopardize future funding.
With the federal government pouring billions into infrastructure through the IIJA, expectations for speed, transparency, and accountability have never been higher. Agencies that modernize their risk review process, at both the project and program level, will be far better equipped to meet those expectations.
So, if your agency is preparing for a wave of capital delivery or already navigating it, now is the time to reassess whether traditional reviews are still enough because when delivery risk is growing, confidence shouldn’t be falling.

From AI-driven forecasting to rapid decision making: why we built Driving Paths
In the past five years at nPlan I have been lucky enough to work on the deployment of ML-powered products in an industry that has seen little adoption of AI technologies in general: management of large capital projects. In this post I would like to share one success story and why I think it worked.

Why AI is more accurate than humans at probabilistic risk forecasting
Alan Mosca and Georgia Stillwell explore how machine learning surpasses human expertise in risk forecasting. They discuss the limitations of human judgment in traditional risk management, emphasizing overconfidence and inaccuracy in estimating probabilities.





