Scanning a building is the easy part. Turning that data into a model that coordinates disciplines, survives handover and stays current is where AEC projects lose time and money.
This session follows four points along that path. It opens with the standards themselves, and what is missing from the way existing conditions documentation gets specified, since deliverables that technically meet spec still miss expectations and leave contractors with models they cannot trust. From there, practitioners weigh how to choose among handheld, 360, UAS, mobile and terrestrial capture for a given project, and what it costs when the choice is wrong. AI point cloud segmentation then takes on the manual modeling bottleneck, feeding scan to BIM, quantity takeoff and facility management from a single dataset. The hour closes with a million square foot project delivered across 18 working models and more than 1,500 sheets, and the point at which BIM as a primary deliverable started working against the team. Attendees will leave with a clearer view of which steps can be automated, which cannot, and where the workflow tends to fail.
This session will include the following presentations:
What’s Missing from LOA and LOD in Specifying Existing Conditions Documentation?
Presented by John Russo, U.S. Institute of Building Documentation (USIBD)
For years, the industry has relied on two standards to specify existing conditions documentation: the USIBD’s Level of Accuracy (LOA) and the BIM Forum’s Level of Development (LOD). Both are widely adopted. Both are genuinely useful. And yet, projects still go sideways. Owners receive deliverables that technically meet spec but miss expectations, service providers over-model or under-deliver, contractors inherit models they can’t trust, and LOD 500 remains the most cited and least understood designation in the specification. Meanwhile, “Level of Development” and “Level of Detail” continue to be used interchangeably, creating confusion that ripples through contracts, scopes, and deliverables.
The problem isn’t that LOA and LOD are wrong, it’s that a stool doesn’t stand on two legs. Something fundamental appears to be missing from how we specify existing conditions, and a concept now taking shape within the USIBD, in coordination with the BIM Forum, may be the third leg the framework has always needed.
This session offers attendees a first look at that emerging concept before it becomes a published specification and invites the audience to help shape it. We’ll examine where the current two-standard system breaks down for each stakeholder (owner, design professional, contractor, and service provider), share best practices for using LOA and LOD today, and preview how the proposed third element could finally give LOD 500 real meaning and resolve the Development-vs-Detail confusion for good. Attendee feedback from this session will directly inform the concept’s development.
Right-Sizing Reality Capture for Infrastructure Beyond Lidar
Presented by Aaron Stock, HNTB
Infrastructure owners have more reality capture technologies available than ever before, yet selecting the appropriate solution for a specific project remains a significant challenge. Handheld image-based capture, 360-degree imagery, UAS photogrammetry, mobile mapping systems, terrestrial laser scanning, and other emerging technologies each offer unique advantages depending on project objectives, site conditions, required accuracy, and available resources.
This practitioner-led session will explore how HNTB evaluates and applies a range of reality capture solutions to support infrastructure planning, asset management, inspection, engineering review, and digital delivery initiatives. Through multiple real-world transportation and infrastructure examples, attendees will learn how different technologies—including the Looq AI Platform, UAS-based photogrammetry, 360° reality capture, mobile mapping, and laser scanning—can be used independently or in combination to document existing conditions, reduce field time, improve safety, and create reusable digital records that support project and asset lifecycle objectives.
The presentation will compare practical considerations including project constraints, data requirements, site accessibility, operational impacts, processing workflows, and end-use applications. Attendees will gain insight into how infrastructure owners and consultants can develop fit-for-purpose reality capture strategies that maximize value while avoiding unnecessary complexity or cost.
Unleashing AI Point Cloud Segmentation for BIM, QTO, and FM
Presented by Wai Yeung Yan, Nova Spatium Inc.
As 3D laser scanning becomes standard practice in the AEC industry, professionals face a massive bottleneck in turning billions of raw spatial points into structured, actionable intelligence. Hand-modeling objects, manually calculating quantities, and updating facility records remains incredibly labor-intensive and prone to human error. This session explores how advanced AI-driven point cloud segmentation is breaking this bottleneck, demonstrating how machine learning can automatically classify, segment, and extract key features from raw scan data to bypass hours of manual labor.
We will walk through the entire lifecycle of a point cloud, demonstrating how a single AI-powered dataset feeds directly into three critical industry workflows. First, we will examine how this technology accelerates the Scan-to-BIM process to create highly accurate, parametric digital twins. Next, we will explore how it instantly extracts material and asset quantities for precise Quantity Takeoffs (QTO) to streamline cost estimation. Finally, we will address how it populates asset databases with real-world spatial intelligence to optimize long-term Facility Management (FM) and operations.
Through real-world case studies, we will compare traditional manual workflows with AI-assisted pipelines, highlighting the massive time savings, accuracy improvements, and return on investment. Attendees will walk away with a deep understanding of how AI algorithms classify complex spatial data, practical strategies to integrate automated feature extraction into existing modeling pipelines, and actionable methods to bridge the gap between raw geospatial scans and long-term facility operations.
Navigating the Limitations of Technology
Presented by Brian Strydom, Cameron Clark, and Daniel Barsky, Urbahn Architects
This presentation will examine the delivery of a 1.1 million square foot project that required the project team to rethink how BIM, data management, documentation production, automation, and quality control operate at scale. With more than 1,500 sheets distributed across four documentation models and 18 working models, the project created an environment where standard workflows had to be tested, refined, automated, and managed with a high degree of consistency.
The session will discuss the advantages and limitations of using BIM as a primary deliverable, including the benefits of detailed modeling for coordination, clash analysis, and VR walkthroughs. We will also dive into the technical challenges created by large files with complex geometry, long synchronization times, and performance constraints. Attendees will gain insight into the practical strategies used to balance modeling accuracy with usability and project efficiency.
The role of structured data extraction, tracking, and analysis in managing project complexity will be outlined. Topics will include ceiling matrix coordination, per-sheet submission tracking, automated data workflows, AI-assisted production, and the management of more than 169,000 documentation data points used to support decision-making and consistency.
A major area of focus was developing the right BIM management framework required for a project of this scale. This framework includes weekly model QA, view quality control, workset strategy, model-splitting workflows, hardware planning, documentation standards, Revit etiquette, automation, proactive communication, and the importance of clear submission goals. We will also discuss how staffing strategies, document audit trails, and dedicated model ownership helped maintain productivity across multiple models while preserving consistency and accountability.
We will explore publishing and printing workflows for large drawing sets, including alternating publishing schedules, sheet-by-sheet progress prints, batch printing automation, and the practical limitations of both hardware and Revit.
Finally, the presentation will show how AI assistants and automation tools helped accelerate repetitive QA and data-entry tasks, adapt workflows to unique project conditions, and support weekly clash detection.
