Assess an existing, market-available AI tool for DoD medical-facility delivery. First gate it for DoD feasibility (Tier 1–3), then score it 1–5 across five weighted criteria for a 0–100 result. Work top to bottom, or turn on Guided mode (top-right) for a step-by-step walkthrough — and open any “How to fill this section” bar for help.
Tool Info & Scoring
Tier --
1
Tool basics
Tool Name — the specific commercial product & vendor (e.g., "Autodesk Forma"). Evaluate existing tools, not concepts.
Lifecycle Phase — where it primarily applies: Planning & Programming, Design (BIM/VDC), Construction, or Operations & Lifecycle.
Choose the phase of greatest impact; note any secondary or cross-cutting use in the SME comments.
2
AI Category Classification
Pick what the tool's AI mainly does. Tap ? for what each category means.
Pick the Primary Category for the tool's core AI capability — this complements the lifecycle phase (required).
Add a Secondary Category only if it genuinely spans two capabilities — otherwise leave "None".
"AI" means real ML, computer vision, NLP, generative, or predictive capability — not marketing claims. Hover a dropdown option for examples.
3
Tier Gate (Feasibility)
Answer Yes / Partial / No from evidence — this decides the tier before scoring. Tap ? for what each question means.
This is a feasibility gate, applied before scoring so infeasible tools aren't elevated by features alone.
Enterprise support (built for regulated orgs) · DoD-compatible deployment (cloud / on-prem / FedRAMP) · basic security (access, encryption, logging) · government precedent.
Any No → Tier 3. Any Partial (no No) → Tier 2. Two+ Yes → Tier 1. Base answers on evidence, not impressions.
Enterprise / regulated environment support?
Deployment compatible with DoD?
Basic security controls?
Precedent in government contexts?
Tier 1 if no "No" and 2+ "Yes" | Tier 2 if any "Partial" | Tier 3 if any "No"
4
Weighted Scoring Rubric
Score 1 (poor) to 5 (excellent). Tap ? by any category for exactly what to look for.
Score each category 1 (poor) → 5 (excellent) after tiering; half-steps (e.g. 3.5) are allowed.
Weights favor the mission — DoD feasibility & UFC relevance over novelty: Feasibility 30% · UFC 25% · Interop 20% · Maturity 15% · ROI 10%.
Under 3.0 on Feasibility, UFC, or Interop raises a non-disqualifying soft flag for SME attention.
Soft thresholds: < 3.0 triggers flags for Feasibility, UFC, or Interop.
Evaluation Review Dashboard
0
Total Tools
0
Tier 1 (DoD-Ready)
0
Tier 2 (Possible)
0
Tier 3 (Not Feasible)
Tool Name ▲
Lifecycle ▲
AI Category ?6 AI capability types: Generative Design, QA/QC, Cost & Schedule, Reality Capture, Digital Twins, and Document Intelligence. Classifies what kind of AI the tool provides.▲
Score ?Weighted 0-100 score across 5 categories: DoD Feasibility: 30% UFC Relevance: 25% Interoperability: 20% Maturity: 15% ROI Potential: 10% 80+ High, 70-79 Strong, 60-69 Moderate, <60 Low▲
SME Status ?Subject Matter Expert validation status. Scores are initially assigned by the research team, then validated by MCX, ERDC, BIM/VDC, or Cybersecurity SMEs at 35%, 65%, and 95% milestones.▲
Date ▲
Actions
No evaluations saved yet.
Saved Evaluations Local DB
0 entries
No saved evaluations yet. Complete an evaluation and click "Save to Database".
Reports
Task #5 milestone submittals for the AI Tooling Evaluation Framework (UFC 4-510-01). Open a completed report to read it in full.
100%
Final Submittal
Coming soon
Task #5 – AI Tooling Evaluation Framework
35% Concept Research Assessment Report
Executive Summary
This report documents the 35% Concept Research Assessment for Task #5, focused on establishing a structured, defensible framework for evaluating Artificial Intelligence (AI) tools applicable to the planning, design, construction, and lifecycle management of Department of Defense (DoD) medical facilities. The intent of Task #5 is not to develop new AI solutions, but to systematically assess existing, market-available tools to determine their feasibility within DoD constraints and their potential impact on UFC 4-510-01 criteria and associated workflows.
At the 35% milestone, the primary objective is to demonstrate methodology rather than reach conclusions. This phase establishes the evaluation boundaries, a lifecycle-based AI classification taxonomy, a tiered feasibility model, and a weighted scoring governance structure. Together, these components create a repeatable, auditable approach that aligns with USACE and DHA expectations, prioritizes cybersecurity and standards compliance, and avoids premature endorsement of immature technologies.
Key outcomes of the 35% effort include:
A clear, defensible definition of “AI” as it applies to Task #5, distinguishing operationally relevant tools from marketing-driven claims.
A three-tier feasibility classification system (DoD-Ready, DoD-Possible, Not Currently Feasible) that reflects real-world deployment constraints rather than binary approval decisions.
A lifecycle-first AI taxonomy aligned with medical facility delivery phases, with built-in flexibility to accommodate cross-cutting capabilities.
A weighted scoring rubric that prioritizes DoD feasibility and UFC relevance over novelty, ensuring alignment with the project’s criteria-driven mission.
Defined evaluation outputs, soft thresholds, and governance logic to support transparency, consistency, and SME-informed judgment.
This framework positions the project for the 65% phase, where the evaluation matrix will be populated with a broader set of tools and preliminary shortlisting logic will be applied. The 35% framework ensures continuity, future-proofing, and credibility as findings mature and inform potential updates to UFC 4-510-01.
1. Introduction and Purpose
Task #5 provides MCX with a structured methodology to evaluate AI tools as standards-ready capabilities rather than experimental technologies. The focus is on identifying tools that can measurably improve outcomes in DoD medical facility delivery while remaining compatible with USACE workflows, cybersecurity requirements, and UFC governance.
The 35% submittal is intended to:
Demonstrate the proposed evaluation framework and documentation format.
Confirm the completeness and appropriateness of AI categories.
Establish a clear path forward for comparative tool evaluation at later milestones.
This phase does not produce recommendations for adoption. Instead, it lays the analytical foundation required for defensible decision-making later in the project.
2. Evaluation Boundaries and Guiding Principles
2.1 Mission Statement
Task #5 evaluates AI tools to determine what exists today, what is usable within DoD constraints, and how those capabilities could inform future UFC criteria and workflows. The task explicitly avoids AI development or pilot implementation.
2.2 Guiding Principles
Five principles guide all evaluation decisions:
Standards First — AI must support and clarify UFC criteria, not replace them.
Evidence Over Hype — Preference is given to tools with demonstrated value and documented use.
DoD Feasibility — Deployment model, cybersecurity posture, and data handling are decisive factors.
Criteria Impact — Evaluation consistently asks how a tool could influence UFC 4-510-01.
Future-Proofing — The framework supports structured, phased adoption rather than one-off pilots.
2.3 Definition of AI for Task #5
For this research, AI tools are defined as software platforms or systems that use machine learning, computer vision, natural language processing, generative algorithms, or predictive analytics to automate, augment, or materially improve decision-making across the medical facility lifecycle. This definition is intentionally broad yet excludes tools that rely solely on marketing claims without functional AI capability.
3. Tiered Feasibility Classification
Rather than a binary approved/not-approved approach, tools are classified into one of three feasibility tiers:
Tier 1 — DoD-Ready (Near-Term Feasible): Enterprise-grade tools with documented use in regulated environments and plausible operation within DoD cybersecurity and data constraints.
Tier 2 — DoD-Possible (Mid-Term Feasible): Mature commercial tools that provide value but require policy, IT, or contractual changes for DoD deployment.
Tier 3 — Not Currently Feasible (Long-Term / Informational): Early-stage or research-driven tools lacking enterprise controls or requiring unrestricted cloud access.
Tiering functions as a feasibility gate and is applied before any comparative scoring. This prevents infeasible tools from being artificially elevated due to feature richness alone.
4. AI Classification Strategy
4.1 Lifecycle-Based Taxonomy
AI tools are initially organized by primary medical facility lifecycle phase:
Planning & Programming
Design (BIM / VDC / Optimization)
Construction
Operations & Lifecycle Management
This structure mirrors UFC organization and USACE workflows, improving interpretability for reviewers and working groups.
4.2 Built-In Flexibility
Each tool may also be tagged with secondary lifecycle applicability and meaningfully cross-cutting capabilities. This allows the framework to evolve at later phases without invalidating earlier work or restructuring prior deliverables.
5. Weighted Scoring Governance
5.1 Scoring Philosophy
Not all evaluation criteria are equally important. The scoring framework is intentionally weighted toward DoD feasibility and UFC relevance, reflecting the project’s standards-driven mission.
5.2 Scoring Categories and Weights
Tools are scored on a 1–5 scale across five categories:
DoD / USACE Feasibility — 30%
UFC & Standards Relevance — 25%
Interoperability & Workflow Integration — 20%
Maturity & Market Adoption — 15%
ROI Potential (Time / Cost / Risk) – 10%
Scores are multiplied by category weights to produce a total weighted score out of 100. Scoring occurs only after feasibility tiering.
6. Thresholds, Outputs, and Governance
6.1 Evaluation Outputs
Each evaluated tool will produce:
Core identification data and lifecycle classification
Feasibility tier assignment
Quantitative weighted scorecard
Qualitative narrative addressing strengths, limitations, assumptions, and medical relevance
Soft thresholds are used to flag risks (e.g., cybersecurity, UFC relevance, interoperability) without disqualifying tools. Final judgments are informed by SME review rather than scores alone.
6.3 SME Validation
Subject Matter Experts review tier assignments, flags, and narratives at the 35%, 65%, and 95% milestones. All adjustments are documented to preserve transparency and auditability.
7. Path Forward to 65%
The 35% framework establishes the methodological backbone for subsequent phases. At 65%, the evaluation matrix will be expanded with additional tools, comparative benchmarking will begin, and preliminary shortlisting logic will be applied to distinguish pilot candidates, criteria-informing tools, and informational references.
8. Conclusion
The 35% Concept Research Assessment successfully establishes a clear, defensible, and future-proof framework for evaluating AI tools relevant to DoD medical facilities. By separating feasibility from
UFC 4-510-01 Update Research Project
Artificial Intelligence Review
(65% Submittal)
USACE Medical Facilities Mandatory Center of Expertise & Standardization
Rogers, Lovelock & Fritz, Inc.
Executive Summary
This report documents an established structured, defensible framework for evaluating Artificial Intelligence (AI) tools applicable to the planning, design, construction, and lifecycle management of Department of Defense (DoD) medical facilities. The intent is not to develop new AI solutions, but to systematically assess existing, market-available tools to determine their feasibility within DoD constraints and their potential impact on UFC 4-510-01 criteria and associated workflows. This framework addresses a key gap: the absence of a standardized, defensible method for evaluating AI tools within DoD constraints, where cybersecurity, interoperability, and standards alignment are critical.
Key outcomes of the effort thus far include:
A clear, defensible definition of “AI” as it applies to this effort, distinguishing operationally relevant tools from marketing-driven claims.
A three-tier feasibility classification system (DoD-Ready, DoD-Possible, Not Currently Feasible) that reflects real-world deployment constraints rather than binary approval decisions.
A lifecycle-first AI taxonomy aligned with medical facility delivery phases, with built-in flexibility to accommodate cross-cutting capabilities.
A weighted scoring rubric that prioritizes DoD feasibility and UFC relevance over novelty, ensuring alignment with the project’s criteria-driven mission.
Defined evaluation outputs, soft thresholds, and governance logic to support transparency, consistency, and SME-informed judgment.
This framework positions the project, where the evaluation matrix is populated with a broad set of tools and applied shortlisting logic. The framework ensures continuity, future-proofing, and credibility as findings mature and inform potential updates to UFC 4-510-01.
Participants
The following individuals contributed to the content of this report:
Michael Lanier, PMP, Project Integrator Team Lead, USACE Medical Facilities MCX
Allison Pride, RA, Senior Architect, USACE Medical Facilities MCX
Van Woods, Senior Researcher, USACE ERDC Information Technology Laboratory
Brian R. White, AIA NCARB LEED AP, Architect, RLF Architecture Engineering Interiors
1. Introduction and Purpose
This research effort provides a structured methodology to evaluate AI tools as standards-ready capabilities rather than experimental technologies. The focus is on identifying tools that can measurably improve outcomes in DoD medical facility delivery while remaining compatible with USACE workflows, cybersecurity requirements, and UFC governance.
2. Evaluation Boundaries and Guiding Principles
2.1 Mission Statement
This research effort evaluates AI tools to determine what exists today, what is usable within DoD constraints, and how those capabilities could inform future UFC criteria and workflows. The task explicitly avoids AI development or pilot implementation.
2.2 Guiding Principles
Five principles guide all evaluation decisions:
Standards First — AI must support and clarify UFC criteria, not replace them.
Evidence Over Hype — Preference is given to tools with demonstrated value and documented use.
DoD Feasibility — Deployment model, cybersecurity posture, and data handling are decisive factors.
Criteria Impact — Evaluation consistently asks how a tool could influence UFC 4-510-01.
Future-Proofing — The framework supports structured, phased adoption rather than one-off pilots.
2.3 Definition of AI
For this research, AI tools are defined as software platforms or systems that use machine learning, computer vision, natural language processing, generative algorithms, or predictive analytics to automate, augment, or materially improve decision-making across the medical facility lifecycle. This definition is intentionally broad yet excludes tools that rely solely on marketing claims without functional AI capability.
3. Tiered Feasibility Classification
Rather than a binary approved/not-approved approach, tools are classified into one of three feasibility tiers:
Tier 1 — DoD-Ready (Near-Term Feasible): Enterprise-grade tools with documented use in regulated environments and plausible operation within DoD cybersecurity and data constraints.
Tier 2 — DoD-Possible (Mid-Term Feasible): Mature commercial tools that provide value but require policy, IT, or contractual changes for DoD deployment.
Tier 3 — Not Currently Feasible (Long-Term / Informational): Early-stage or research-driven tools lacking enterprise controls or requiring unrestricted cloud access.
Tiering functions as a feasibility gate and is applied before any comparative scoring. This prevents infeasible tools from being artificially elevated due to feature richness alone. Only tools classified as Tier 1 (DoD-Ready) or Tier 2 (DoD-Possible) proceed to weighted scoring, ensuring that infeasible tools are excluded from comparative evaluation.
4. AI Classification Strategy
4.1 Lifecycle-Based Taxonomy
AI tools are initially organized by primary medical facility lifecycle phase:
Planning & Programming
Design (BIM / VDC / Optimization)
Construction
Operations & Lifecycle Management
This structure mirrors UFC organization and USACE workflows, improving interpretability for reviewers and working groups.
4.2 Built-In Flexibility
Each tool may also be tagged with secondary lifecycle applicability and meaningfully cross-cutting capabilities. This allows the framework to evolve at later phases without invalidating earlier work or restructuring prior deliverables.
5. Weighted Scoring Governance
5.1 Scoring Philosophy
Not all evaluation criteria are equally important. The scoring framework is intentionally weighted toward DoD feasibility and UFC relevance, reflecting the project’s standards-driven mission.
5.2 Scoring Categories and Weights
Tools are scored on a 1–5 scale across five categories:
DoD / USACE Feasibility — 30%
UFC & Standards Relevance — 25%
Interoperability & Workflow Integration — 20%
Maturity & Market Adoption — 15%
ROI Potential (Time / Cost / Risk) – 10%
Scores are multiplied by category weights to produce a total weighted score out of 100. Scoring occurs only after feasibility tiering.
6. Thresholds, Outputs, and Governance
6.1 Evaluation Outputs
Each evaluated tool will produce:
Core identification data and lifecycle classification
Feasibility tier assignment
Quantitative weighted scorecard
Qualitative narrative addressing strengths, limitations, assumptions, and medical relevance
Soft thresholds function as risk indicators (e.g., cybersecurity gaps, low interoperability), prompting additional SME review rather than automatic exclusion. Final judgments are informed by SME review rather than scores alone.
6.3 SME Validation
Subject Matter Experts review tier assignments, flags, and narratives. All adjustments are documented to preserve transparency and auditability.
7. AI Generated Software for Mission-Specific Design Tools
7.1 Context
The Department of Defense relies on highly specialized standards such as UFC 4-510-01 for Military Medical Facilities to ensure consistency, safety, and operational performance in facility design. Updating and implementing these standards requires significant coordination across architects, engineers, and subject-matter experts.
One challenge identified during the early research process is the lack of custom digital tools that translate UFC criteria into practical workflows during planning and design.
Historically, developing these tools required dedicated software development resources or reliance on commercial software platforms that may not align precisely with DoD requirements.
7.2 Emerging Capability: AI-Generated Software
Recent advances in AI coding agents, such as Claude Code, demonstrate a new capability: the ability to generate functional software tools directly from written instructions.
Instead of commissioning a development team, users can describe the desired functionality in natural language and the AI can generate the application structure, logic, and interface.
This enables rapid creation of mission-specific digital tools tailored to unique workflows, including those associated with UFC implementation.
7.3 Demonstration Example
During the research effort supporting this project, a custom AI Evaluation Matrix application was created using Claude Code.
The tool was developed to support the systematic evaluation of artificial intelligence technologies relevant to the design and lifecycle performance of DoD medical facilities.
Rather than relying on existing software platforms, the application was generated through AI-assisted development based on a clear description of required capabilities, including:
Scoring frameworks for AI tools
Tier-gate feasibility screening
Compatibility checks with DoD cybersecurity and deployment constraints
UFC touchpoint mapping to design workflows
The development process required hours rather than weeks or months, demonstrating how AI coding agents can rapidly generate custom research and evaluation tools.
7.4 Relevance to UFC Research Objectives
This example highlights how AI-generated software can support DoD research and design processes in several ways:
Rapid prototyping of specialized tools
Teams can quickly create digital tools that interpret and apply UFC criteria within design workflows.
Customization for government standards
Tools can be tailored specifically to UFC, USACE, and DHA requirements rather than adapting commercial software designed for general markets.
Accelerated innovation cycles
Ideas for improving design evaluation or compliance verification can be implemented and tested almost immediately.
7.5 Strategic Implications for DoD Design
The ability to rapidly generate software tools may significantly change how design and research teams support military facility standards.
Rather than waiting for commercial software vendors to develop new capabilities, teams may increasingly be able to create targeted applications on demand to support planning, design evaluation, and facility lifecycle analysis.
This capability aligns with the broader goal of the research task:
to identify emerging AI technologies that measurably improve the design, construction, and lifecycle performance of DoD medical facilities while remaining compatible with USACE and DHA standards.
7.6 Key Takeaway
AI coding agents demonstrate that custom digital tools for implementing UFC guidance can now be created quickly and cost-effectively.
The example of the AI Evaluation Matrix illustrates how AI-generated software can enable new workflows that support DoD design research, analysis, and decision-making.
As these capabilities mature, they may allow research teams and design organizations to rapidly develop mission-specific applications that improve the implementation of UFC standards across the entire facility lifecycle.
Examples of custom AI Evaluation Matrix application was created using Claude Code.
8. AI Evaluation Matrix
“Screen shots of matrix”
9. Conclusion
This research effort successfully establishes a clear, defensible, and future-proof framework for evaluating AI tools relevant to DoD medical facilities. By separating feasibility from performance-based scoring, the framework ensures that only viable, DoD-aligned tools are evaluated in depth, supporting transparent, defensible, and criteria-driven decision-making for future UFC updates.
Appendix A – USACE AI Systems
Tools Currently Available
The suite of tools available for general use by USACE employees continues to evolve at a rapid pace. Here is a snapshot of the currently available tools:
Generative Artificial Intelligence (GenAI) platforms like ChatGPT, GROK, Google Gemini, Copilot, and Meta AI that leverage large language models (LLMs) offer transformative benefits in terms of productivity, automation, and innovation. However, their use presents significant security, operational, and ethical risks that must be considered in accordance with USACE, Army, and DoD guidance. The USACE AI guidance provided below has been developed in alignment with guidance provided by the Army CIO in CS-SEC-SC-003 (CUI) Safeguarding United States Government Data and Information on Open Source and Commercial Applications and Services and ADS-GOV-AI-024 Chief Information Officer Guidance on Generative Artificial Intelligence and Large Language Models.
USACE is actively working on strategy, policies, and the deployment of AI tools to the workforce to include leveraging DoD and Army provisioned AI tools such as CamoGPT, Vantage, NIPRGPT, and the Army Enterprise LLM Workspace (AskSage). Future deployments of Microsoft CoPilot with CUI / IL5 authorization across our USACE Microsoft 365 platform are coming this Fall. Many other updates are anticipated through the end of 2025 with new DoD, Army, and USACE capabilities that are currently in development. We will distribute updates and recommended training on these capabilities as they are deployed.
Please see the below USACE guidance when using AI platforms as part of your work.
Currently Approved GenAI Platforms for CUI or Sensitive Data (these require CAC authentication):
General GenAI / LLM Tools:
CamoGPT: https://camogpt.army.mil/
Army Enterprise LLM Workspace (AskSage) : https://chat.genai.army.mil/
Initial trial tokens available. Army is working to provision additional resources and USACE will be developing procurement guidance if necessary.
NIPRGPT: https://niprgpt.mil/
Powerful Data Analytics and AI Platforms:
Army Vantage Data Analytics Platform: https://vantage.army.mil/
ADVANA Analytics: https://advana.data.mil/
Contracting / Acquisition Focused Tools:
AcqBot: https://acqbot.niprgpt.mil/
CDAO Tradewinds: https://www.tradewindai.com/
NOTE: Other LLM / GenAI platforms such as public versions of ChatGPT, Copilot, GROK, Claude, Meta AI available through commercial vendors are NOT authorized for any sensitive or CUI information and should only be used for Distribution A, non-sensitive activities.
How Can You Learn More?
For more information on USACE AI efforts and links to training, prompt engineering guidance, and more, reach out to the USACE AI Community of Practice: https://usace.dps.mil/sites/KMP-AI
A wealth of free training on GenAI tools and prompt engineering is also available through the Army’s Udemy platform at: https://armyciv.udemy.com/
Approved Usage Guidelines:
Do not input Classified, CUI, or sensitive data (e.g., PII, HIPAA, procurement sensitive, attorney-client privileged, and pre-decisional materials) into publicly available GenAI platforms. Please review the examples provided in the attached reference Example CUI and Sensitive Information Types and Media Reported Security Failures.
Mandatory CUI Training: https://securityawareness.dcsa.mil/cui/index.html
Only use approved, authorized or USACE hosted GenAI platforms when working with sensitive data, ensuring the platform’s impact level is appropriate for the data classification.
Public GenAI platforms may only be used with unclassified, non-sensitive, non-CUI data.
Aggregation of multiple data points or types of information may still result in CUI output so proceed with caution.
Certain platforms including DeepSeek are explicitly prohibited for Army use under any circumstances.
Validate and verify GenAI generated output prior to usage. Incorporating it into official USACE documents, briefings, or decisions given the abundance of AI hallucinations is a significant issue, including hallucinations that would be considered CUI.
Uploading Files into GenAI Platforms:
Uploading of files into GenAI platforms that are not authorized to handle CUI is currently blocked (such as ChatGPT and Google Gemini).
CIO/G6 is currently re-examining this policy to allow for users to upload non-sensitive, Distribution A information as needed into platforms like ChatGPT to support your work. Be on the look out for updates on this topic.
It is your responsibility to ensure that no CUI files, text, or data are uploaded into unapproved AI platforms that are not authorized to handle CUI.
Information Security and Data Protection:
Comply with USACE cybersecurity policies, Army and DoD security requirements (e.g., DoDI 5200.48 – Controlled Unclassified Information).
Clearly label content that is generated in whole or in part using GenAI tools.
Do not use any sensitive data to train or fine tune GenAI models unless approved by the appropriate data owners.
Be aware that public GenAI platforms retain user inputs and uses the input to train their models, increasing the risk of unauthorized data disclosure. Although specifics of information or user inputs are typically concealed, all information input into GenAI platforms trains these models to generate new information.
Verification and Compliance:
Carefully review GenAI output for accuracy, appropriateness, and alignment with official guidance.
When in doubt, consult your local USACE Security office or CIO/G6 cybersecurity team.
If GenAI outputs will be used in public facing or decision-making documents, always validate with authoritative sources.
For external communication in particular, ensure coordination with appropriate Public Affairs Office personnel who have also implemented Army guidance on AI in public affairs and communications. Please see the attached Army AI Guidance for Public Affairs for more information.
Need Help or Have Questions?
Please coordinate with your USACE CIO/G6 office or contact our Cybersecurity Team DLL-CEIT-ZS-GovernancePolicy@usace.army.mil for clarification on approved usage and platform availability.
Please contact Governance & Architecture Division for any IT policy related questions USACE-CIO-GAD@usace.army.mil
Contact PrivacyOffice@usace.army.mil for any privacy related question.
For questions on AI and communications, engagement with your local Public Affairs Office.
Why It Matters:
Proper use of GenAI protects Army data, USACE missions, and public trust. As stewards of critical information of many forms, it’s essential we strike the right balance between modernization using AI and operational security.
References / Attachments:
CS-SEC-SC-003 (CUI) Safeguarding United States Government Data and Information on Open Source and Commercial Applications and Services.
ADS-GOV-AI-024 Chief Information Officer Guidance on Generative Artificial Intelligence and Large Language Models
Example CUI and Sensitive Information Types and Media Reported Security Failures
Army AI Guidance for Public Affairs
Thank you for your continued commitment to safeguarding information as we responsibly integrate emerging technologies into USACE operations.
EA: Ms. Jocelyn Johnson / 202.579.1024 / Jocelyn.N.Johnson@usace.army.mil
Appendix C – USACE AI Adoption Strategy
U.S. ARMY CORPS OF ENGINEERS (USACE)
ARTIFICIAL INTELLIGENCE (AI) ADOPTION
STRATEGY
6 February 2025
This appendix provides a comprehensive A–Z glossary of terms relevant to this research task (Artificial Intelligence Review) supporting the update of UFC 4-510-01 (Design: Military Medical Facilities). Definitions reflect DoD, USACE, AEC, cybersecurity, BIM, and AI governance contexts. This is a living appendix and will be updated as additional tools, platforms, and workflows are evaluated.
A
Algorithm — A defined computational procedure used to process data and generate outputs.
AEC — Architecture, Engineering, and Construction industry sector.
API (Application Programming Interface) — Mechanism allowing software systems to exchange data and functionality.
Asset Management — Lifecycle tracking and optimization of facility systems and equipment.
Audit Log — Recorded system/user activity for accountability and compliance.
B
BIM (Building Information Modeling) — Digital representation of facility physical and functional characteristics.
BEP (BIM Execution Plan) — Document defining BIM uses, standards, and responsibilities.
Black Box Model — AI system with limited transparency into internal decision logic.
Building Automation System (BAS) — Control system managing HVAC, lighting, and other building systems.
C
Change Order — Contract modification affecting scope, cost, or schedule.
Clash Detection — Identification of model conflicts (e.g., MEP vs structure).
Cloud Computing — Vendor-hosted computing resources accessed via network.
COBie — Structured data format for facility asset handover.
Computer Vision (CV) — AI enabling interpretation of images and video.
Constructability — Practical buildability of a design.
Cybersecurity Posture — Overall security readiness of a system or vendor.
D
Data Governance — Policies and controls managing data quality, security, and usage.
Data Residency — Geographic/legal location of stored data.
Data Sensitivity — Risk classification of data (e.g., PHI, drawings, mission data).
Digital Delivery — End-to-end digital project information workflow.
Digital Twin — Dynamic digital representation of a facility linked to performance data.
E
Edge Computing — Localized processing near data source.
Enterprise-Grade — Software designed for large regulated organizations.
Explainability — Ability to interpret and justify AI outputs.
F
FedRAMP — Federal authorization program for cloud security compliance.
Feasibility Tier — Classification of AI deployability (Tier 1 DoD-Ready; Tier 2 DoD-Possible; Tier 3 Informational).
Facility Lifecycle — Planning, design, construction, operations, and sustainment phases.
G
Generative AI — AI capable of producing designs, layouts, text, or alternatives.
Generative Design — Algorithm-driven creation of optimized design options.
H
Handover Data — Information transferred from construction to operations.
Hybrid Deployment — Combination of cloud and on-premise computing.
I
IFC (Industry Foundation Classes) — Open BIM data exchange standard.
Interoperability — Ability of systems to exchange and use shared data.
Integration Risk — Potential workflow disruption due to poor system compatibility.
J
Justification Narrative — Documented reasoning supporting evaluation scores or recommendations.
K
Knowledge Model — Structured representation of rules or expertise within software.
Soft Threshold Flag — Non-disqualifying risk indicator in evaluation framework.
Standards Relevance — Degree of impact on UFC criteria or workflows.
T
Tier 1 (DoD-Ready) — Near-term feasible within DoD constraints.
Tier 2 (DoD-Possible) — Feasible with policy or IT adjustments.
Tier 3 (Informational) — Not currently deployable.
Traceability — Ability to track decisions and data lineage.
U
UFC (Unified Facilities Criteria) — DoD design standards governing military facilities.
UFC Touchpoint — Section or workflow within UFC potentially influenced by AI.
V
Validation — Confirmation that system outputs meet intended requirements.
Vendor Stability — Financial and operational reliability of software provider.
W
Workflow Integration — Alignment of tool functionality with existing project processes.
Weighted Scoring — Assignment of percentage-based importance to evaluation criteria.
X
XML (eXtensible Markup Language) — Structured data format sometimes used in BIM exchanges.
Y
Yield Optimization — Improvement of performance or efficiency through data-driven methods.
Z
Zero Trust Architecture — Cybersecurity model requiring continuous verification of users and devices.
UFC 4-510-01 Update Research Project
Artificial Intelligence Review
(95% Submittal — DRAFT)
USACE Medical Facilities Mandatory Center of Expertise & Standardization
Rogers, Lovelock & Fritz, Inc.
Executive Summary
This report documents the 95% milestone of Task #5, the Artificial Intelligence Review supporting the update of UFC 4-510-01 (Design: Military Medical Facilities). Building directly on the framework established at 35% and populated at 65%, the 95% submittal presents the completed evaluation matrix, the resulting feasibility tiering, and an in-depth assessment of every tool currently evaluated — with particular focus on the feasible, DoD-Ready tools recommended for pilot consideration.
The evaluation methodology is unchanged from prior submittals: a feasibility tier gate is applied before scoring, and only the mission-weighted rubric (DoD/USACE feasibility, UFC relevance, interoperability, maturity, and ROI) produces the final 0–100 score. This preserves a defensible, auditable process in which infeasible tools are not elevated by features alone.
At the 95% milestone the portfolio contains 38 tools. Thirty-one have been fully scored; seven newly identified candidates have been added and are pending scoring for the 100% submittal. Of the 31 scored tools, 8 are classified Tier 1 (DoD-Ready), 19 are Tier 2 (DoD-Possible), and 4 are Tier 3 (Not Currently Feasible). The average weighted score across scored tools is approximately 70 out of 100.
The strongest performers cluster around BIM quality assurance and model checking, healthcare space and requirements data, enterprise construction management, secure generative AI, and facility operations and digital twins. Feasibility — deployment model, cybersecurity posture, and government precedent — remains the decisive differentiator: enterprise vendors with documented government use rise to Tier 1, while capable startups without a clear government deployment path fall to Tier 2 or Tier 3 despite strong feature sets.
Participants
The following individuals contributed to the content of this report:
Michael Lanier, PMP, Project Integrator Team Lead, USACE Medical Facilities MCX
Allison Pride, RA, Senior Architect, USACE Medical Facilities MCX
Van Woods, Senior Researcher, USACE ERDC Information Technology Laboratory
Brian R. White, AIA NCARB LEED AP, Architect, RLF Architecture Engineering Interiors
1. Introduction and Purpose
The 95% submittal advances Task #5 from framework and population to findings and recommendations. Its purpose is to present the completed evaluation matrix, document how each tool performed against the mission-weighted rubric, and identify the feasible tools that warrant near-term pilot consideration or feasibility review.
As in prior phases, this effort does not develop AI or authorize adoption. It provides MCX and the working group with a structured, defensible basis for decision-making and for identifying where AI capabilities could inform future updates to UFC 4-510-01.
This submittal is issued in DRAFT for SME validation. Scores, tier assignments, and narratives remain subject to SME review and refinement prior to the 100% (Final) submittal.
2. Evaluation Approach (Recap)
2.1 Guiding Principles
Five principles continue to guide every evaluation decision:
Standards First — AI must support and clarify UFC criteria, not replace them.
Evidence Over Hype — Preference is given to tools with demonstrated value and documented use.
DoD Feasibility — Deployment model, cybersecurity posture, and data handling are decisive factors.
Criteria Impact — Evaluation consistently asks how a tool could influence UFC 4-510-01.
Future-Proofing — The framework supports structured, phased adoption rather than one-off pilots.
2.2 Definition of AI
AI tools are defined as software platforms that use machine learning, computer vision, natural language processing, generative algorithms, or predictive analytics to automate, augment, or materially improve decision-making across the medical facility lifecycle — excluding tools that rely on marketing claims without functional AI capability.
3. Tiered Feasibility Classification
A feasibility tier gate is applied before scoring. Each tool is assessed on four gate questions — enterprise/regulated support, DoD-compatible deployment, basic security controls, and government precedent — and assigned to one of three tiers:
Tier 1 — DoD-Ready (Near-Term Feasible): Enterprise-grade tools with documented use in regulated environments and plausible operation within DoD cybersecurity and data constraints.
Tier 2 — DoD-Possible (Mid-Term Feasible): Mature commercial tools that provide value but require policy, IT, or contractual changes for DoD deployment.
Tier 3 — Not Currently Feasible (Long-Term / Informational): Tools lacking a clear enterprise or government deployment path, retained for reference and criteria-informing value.
4. Weighted Scoring Model
Tools that clear the gate are scored 1–5 across five mission-weighted categories, producing a total out of 100:
DoD / USACE Feasibility — 30%
UFC & Standards Relevance — 25%
Interoperability & Workflow Integration — 20%
Maturity & Market Adoption — 15%
ROI Potential (Time / Cost / Risk) – 10%
Soft-threshold flags are raised when Feasibility, UFC relevance, or interoperability score below 3.0. Flags are non-disqualifying; they direct SME attention rather than removing a tool.
5. Portfolio Results Overview
5.1 Summary Statistics
The evaluated portfolio contains 38 tools. Thirty-one are fully scored and seven are newly added and pending scoring. Among the scored tools, the average weighted score is approximately 70 out of 100, and scores range from 92.5 (Solibri) to 48.5 (Document Crunch).
5.2 Tier Distribution
Of the 31 scored tools, the feasibility gate produced the following distribution:
Tier 1 — DoD-Ready (8 tools): Solibri, dRofus, Procore Helix, Ask Sage, Honeywell Forge, Siemens Building X, Autodesk Pype (AutoSpecs), and IBM Maximo Application Suite.
Tier 2 — DoD-Possible (19 tools): Revizto, Autodesk Tandem, Autodesk Forma, Newforma Konekt, Hypar, Autodesk Construction IQ, TestFit, ALICE Technologies, Willow, Plannerly, nPlan, Reconstruct, OpenSpace, Digital Blue Foam, Togal.AI, Facilio, Buildots, 75F, and Spot AI.
Tier 3 — Not Currently Feasible (4 tools): SWAPP AI, Finch3D, qbiq, and Document Crunch.
5.3 Score Bands
Eight tools scored in the High band (80+), six in the Strong band (70–79), thirteen in the Moderate band (60–69), and four in the Low band (below 60). A high score does not by itself imply readiness; a tool must clear the feasibility gate and earn SME validation to be recommended as a pilot candidate.
5.4 Results by Lifecycle Phase
Feasible, high-scoring tools appear across all four lifecycle phases: dRofus in Planning & Programming; Solibri and Revizto in Design (BIM/VDC); Procore Helix and Autodesk Pype in Construction; and Ask Sage, Honeywell Forge, Siemens Building X, and IBM Maximo in Operations & Lifecycle. This confirms that AI capability relevant to UFC 4-510-01 is not concentrated in a single phase but distributed across the delivery lifecycle.
6. In-Depth Assessment: DoD-Ready (Tier 1) Tools
The eight Tier 1 tools cleared the feasibility gate on the strength of enterprise support, DoD-compatible deployment, security controls, and documented government or regulated-industry precedent. All are recommended as Pilot Candidates, subject to SME validation.
Solibri — Tier 1, score 92.5 (Feasibility 4.5 · UFC 4.5 · Interop 5 · Maturity 5 · ROI 4). Nemetschek's BIM quality-assurance and rule-based model-checking tool and the top scorer in the portfolio. Its automated checking of clashes, code compliance, and data integrity maps directly to UFC 4-510-01 review and QA workflows, and it deploys inside established Revit/IFC pipelines. Highest recommendation for a design-phase pilot.
dRofus — Tier 1, score 87.5 (Feasibility 4 · UFC 4.5 · Interop 4.5 · Maturity 5 · ROI 4). A data-centric planning and requirements platform purpose-built for healthcare, with room templates keyed to FGI and other standards plus medical equipment and space-program tracking. Its direct alignment with medical-facility programming makes it one of the most UFC-relevant tools evaluated. Strong candidate for a planning-phase pilot.
Procore Helix — Tier 1, score 86.5 (Feasibility 5 · UFC 3.5 · Interop 4 · Maturity 5 · ROI 4). Procore's built-in construction-management AI, earning the portfolio's highest feasibility score alongside top maturity. Its strengths are workflow automation (RFIs, submittals, daily logs) rather than design criteria, so UFC relevance is moderate; nonetheless it is enterprise-grade and construction-phase ready.
Ask Sage — Tier 1, score 84.0 (Feasibility 3 · UFC 5 · Interop 4 · Maturity 5 · ROI 5). A secure generative-AI platform accredited at FedRAMP High and DoD IL5/IL6 — a distinctive authorization posture already fielded across USACE and the Army. It earned the portfolio's highest UFC-relevance and ROI scores; feasibility was scored conservatively given its general-purpose nature, though its existing DoD authorizations are a decisive real-world asset for secure document and requirements analysis.
Honeywell Forge — Tier 1, score 83.5 (Feasibility 5 · UFC 3.5 · Interop 3 · Maturity 5 · ROI 4.5). Honeywell's enterprise operations platform with a building digital twin and energy optimization. Top feasibility and maturity; interoperability is the relative soft spot. Best suited to the operations and lifecycle phase.
Siemens Building X — Tier 1, score 83.5 (Feasibility 4.5 · UFC 3.5 · Interop 4 · Maturity 5 · ROI 4). Siemens' open, vendor-agnostic building-operations platform. Balanced high marks and strong connectivity into mixed facility systems make it a capable operations-phase candidate.
Autodesk Pype (AutoSpecs) — Tier 1, score 80.0 (Feasibility 4.5 · UFC 3.5 · Interop 3.5 · Maturity 4.5 · ROI 4). Autodesk's AI submittal-log and specification-review tool within Autodesk Construction Cloud. Solid across every category and directly relevant to construction-phase compliance and submittal review.
IBM Maximo Application Suite — Tier 1, score 78.5 (Feasibility 4 · UFC 3.5 · Interop 3.5 · Maturity 5 · ROI 4). IBM's enterprise asset-management suite with predictive maintenance and computer-vision inspection. A proven, widely deployed platform for facility and equipment lifecycle management.
Nineteen tools are DoD-Possible: capable, mature products whose government deployment is currently assessed as partial (typically public-cloud SaaS) or whose precedent is not yet clearly governmental. They are treated as criteria-informing inputs — valuable for shaping UFC updates and worth revisiting as government deployment options mature. One, Revizto, is flagged for near-term feasibility review.
Revizto — Tier 2, score 80.0 (High-value; pursue feasibility review). An integrated BIM collaboration and issue-tracking platform with a perfect interoperability score (5). It reached Tier 2 only because government precedent was assessed partial; if a DoD deployment and precedent path is confirmed, it is a strong near-term candidate.
Autodesk Tandem — Tier 2, score 79.5. A digital-twin platform carrying BIM data into operations; strong interoperability (5), with deployment scored partial.
Autodesk Forma — Tier 2, score 77.0. Cloud pre-design and site analysis with strong UFC relevance for early planning; deployment scored partial.
Newforma Konekt — Tier 2, score 76.5. Cloud BIM coordination and issue tracking; capable and mature, with partial deployment/precedent.
Hypar — Tier 2, score 71.5. Generative design and automation with healthcare space-planning focus; relevant but with partial deployment.
Autodesk Construction IQ — Tier 2, score 71.0. Predictive risk analytics within Autodesk Construction Cloud; useful QA/safety signal, partial deployment.
TestFit — Tier 2, score 67.5. Real-time site and feasibility configurator; strong for early options and quantities.
ALICE Technologies — Tier 2, score 67.0. AI construction scheduling and optioneering; strongest in the construction phase.
Willow — Tier 2, score 66.0. Operational digital-twin platform for buildings and portfolios.
Plannerly — Tier 2, score 64.5. BIM requirements planning and checking; a soft feasibility flag was noted.
nPlan — Tier 2, score 64.5. AI schedule-risk forecasting trained on large historical datasets.
Buildots — Tier 2, score 61.5. AI progress tracking via 360° cameras and computer vision.
75F — Tier 2, score 58.0. IoT building automation and HVAC optimization; low band, with an interoperability flag.
Spot AI — Tier 2, score 57.5. Camera-agnostic video intelligence for jobsite safety and security; low band, with soft flags on UFC relevance and interoperability.
8. Not Currently Feasible (Tier 3) — Informational / Parking Lot
Four tools were gated to Tier 3 because a feasibility gate — most often government precedent — was answered "No." Each is retained for reference; several show strong capability and could be reconsidered if enterprise or government deployment and precedent become clear.
SWAPP AI — Tier 3, score 62.0. AI construction-documentation automation with a high interoperability score (4.5); parked pending clearer enterprise/government precedent.
Finch3D — Tier 3, score 61.5. Generative-design copilot for architects; capable but early for DoD deployment.
qbiq — Tier 3, score 57.0. Generative-AI space planning for commercial real estate; strong concept, limited government footing.
Document Crunch — Tier 3, score 48.5. AI contract and specification risk review; useful capability but the lowest overall score, reflecting deployment and precedent gaps.
9. Emerging Candidates for Future Evaluation
Seven additional tools were identified during the 95% effort and added to the matrix; they are pending full scoring for the 100% submittal. Several carry government authorizations that are expected to place them high on feasibility, and two target strong UFC touchpoints. Ask Sage, already scored above, belongs to this same secure-AI category and validates the relevance of this group.
Microsoft 365 Copilot — Pending. Enterprise generative AI available in government clouds (GCC High / DoD) on FedRAMP High and IL5 pathways; USACE has signaled Copilot deployment with CUI/IL5 authorization, suggesting high feasibility.
Azure OpenAI Service — Pending. GPT-4-class models in Azure Government at FedRAMP High / IL5; a high-feasibility foundation for document intelligence and requirements extraction.
Palantir Foundry / AIP — Pending. Defense-grade data and AI platform with extensive DoD/IC accreditation; very high expected feasibility, though less AEC-specific.
ClearEdge3D Verity — Pending. Automated QA/QC and scan-to-BIM verification in Revit/Navisworks; targets UFC quality-verification touchpoints directly.
Trunk Tools — Pending. AI over specs, submittals, RFIs, and contracts; strong UFC-compliance relevance.
Doxel — Pending. Computer-vision construction progress and quality tracking, proven on healthcare and data-center projects.
Augmenta — Pending. Generative AI for MEP and electrical design and routing; relevant to the complex mechanical/electrical systems in medical facilities.
10. Recommendations
Advance the eight Tier 1 tools to scoped pilot or feasibility discussions, subject to SME validation.
Prioritize by phase — dRofus and Solibri for planning and design given their direct UFC alignment; Procore Helix and Autodesk Pype for construction; Ask Sage, Honeywell Forge, Siemens Building X, and IBM Maximo for operations and lifecycle.
Pursue a feasibility review of Revizto to confirm a government deployment and precedent path that could elevate it to Tier 1.
Treat Tier 2 tools as criteria-informing inputs to UFC 4-510-01 updates, and revisit them as government deployment options mature.
Complete scoring of the seven emerging candidates for the 100% submittal; the FedRAMP High / IL5 platforms among them may enter Tier 1.
Maintain SME validation at the 95% and 100% milestones, documenting all tier and score adjustments for auditability.
11. Path to 100% (Final Submittal)
The remaining work to reach 100% is well defined: score the seven emerging candidates, obtain SME validation on the pilot recommendations and tier assignments, finalize UFC touchpoint mapping for the recommended tools, and convert this draft into the final submittal with any adjustments arising from SME review.
12. Conclusion
The 95% submittal delivers a completed, defensible evaluation of the AI-tool portfolio for DoD medical facilities. By applying feasibility gating before mission-weighted scoring, the framework surfaces a credible set of DoD-Ready tools — led by Solibri, dRofus, Procore Helix, and Ask Sage — while retaining a broader set of criteria-informing tools that can shape future UFC updates. The findings position the working group to select pilot candidates with confidence and to complete the Final submittal with SME validation.
Appendix A – Full Results Matrix (Scored Tools)
Tools are listed in descending order of weighted score. Format: score, tier, and recommended path.
Solibri — 92.5, Tier 1, Pilot Candidate.
dRofus — 87.5, Tier 1, Pilot Candidate.
Procore Helix — 86.5, Tier 1, Pilot Candidate.
Ask Sage — 84.0, Tier 1, Pilot Candidate.
Honeywell Forge — 83.5, Tier 1, Pilot Candidate.
Siemens Building X — 83.5, Tier 1, Pilot Candidate.
Appendix B – Emerging Candidates (Pending Scoring)
Microsoft 365 Copilot, Azure OpenAI Service, Palantir Foundry / AIP, ClearEdge3D Verity, Trunk Tools, Doxel, and Augmenta are added to the matrix and will be scored for the 100% submittal.
Help & Support
Everything you need to evaluate AI tools consistently and defensibly for DoD medical facilities — the purpose, the tier gate, the weighted scoring, the team workflow, a glossary, and FAQs. Grounded in the Task #5 AI Tooling Evaluation Framework.
◆ Purpose & guiding principles
Task #5 is a structured, defensible way to evaluate existing, market-available AI tools for the planning, design, construction, and lifecycle of DoD medical facilities. The goal is not to build AI — it's to determine what exists today, what is usable within DoD constraints, and how each tool could inform UFC 4-510-01 criteria and workflows.
Five guiding principles
Standards First
AI must support and clarify UFC criteria, not replace them.
Evidence Over Hype
Prefer tools with demonstrated value and documented use.
DoD Feasibility
Deployment model, cybersecurity posture, and data handling are decisive.
Criteria Impact
Always ask how a tool could influence UFC 4-510-01.
Future-Proofing
Support structured, phased adoption — not one-off pilots.
What counts as "AI" here — software using machine learning, computer vision, natural-language processing, generative algorithms, or predictive analytics to automate, augment, or materially improve decisions across the facility lifecycle. Marketing claims without real AI capability don't qualify.
1 How an evaluation works — 5 steps
1
Tool basics
Name the tool and pick the lifecycle phase where it's mainly used (Planning, Design, Construction, or Operations).
2
AI category
Classify what kind of AI it is — primary, plus an optional secondary (e.g., Generative Design, QA/QC, Digital Twins).
3
Tier gate
Answer 4 Yes / Partial / No feasibility questions. These set the Tier: 1 Ready · 2 Possible · 3 Not feasible.
4
Weighted rubric
Score 5 categories 1–5. Weighted into a 0–100 score (Feasibility 30 · UFC 25 · Interop 20 · Maturity 15 · ROI 10).
5
Review & save
Check the Tier, score, flags, and Recommended Path, add SME notes, then Save — it appears in Dashboard & Gallery.
2 Feasibility tiers & weighted scoring
Every tool is tiered for feasibility first, then scored. Tiering is a gate applied before scoring, so infeasible tools can't be elevated by features alone.
Tier 1 — DoD-Ready
Near-term feasible. Enterprise-grade tools with documented use in regulated environments, plausibly operable within DoD cybersecurity and data constraints. (No "No" and two or more "Yes".)
Tier 2 — DoD-Possible
Mid-term feasible. Mature commercial tools that add value but require policy, IT, or contractual changes to deploy. (Any "Partial", no "No".)
Tier 3 — Not Currently Feasible
Long-term / informational. Early-stage or research-driven tools lacking enterprise controls or requiring unrestricted cloud access. (Any "No".)
Weighted rubric — score 1–5, ×weight → /100
DoD / USACE Feasibility — 30%
Deployment model, cybersecurity posture, and data handling within DoD constraints.
UFC & Standards Relevance — 25%
How directly the tool supports or influences UFC 4-510-01 criteria.
Interoperability & Workflow — 20%
Fit with USACE / BIM workflows and ability to exchange shared data.
Maturity & Market Adoption — 15%
Proven deployment, vendor stability, breadth of use.
ROI Potential — 10%
Measurable time / cost / risk benefit relative to effort.
Reading the result
Score bands
80+ High · 70–79 Strong · 60–69 Moderate · under 60 Low.
Soft flags
A category under 3.0 (Feasibility, UFC, or Interop) is flagged — a non-disqualifying risk indicator for SME attention, never an automatic fail.
Identification + lifecycle, feasibility tier, the weighted scorecard, an SME narrative (strengths, limits, assumptions, medical relevance), and UFC touchpoints.
3 SME validation & milestones
Subject-Matter Experts (MCX, ERDC, BIM/VDC, and Cybersecurity) review tier assignments, soft-threshold flags, and narratives at the 35%, 65%, and 95% milestones. Final judgments are informed by SME review — not scores alone — and every adjustment is documented for transparency and auditability. Use the SME Validation Notes on each evaluation to record who reviewed it, the date, any tier/score changes, and any policy or IT blockers, then set the Validation Status.
4 Team scoring & import
Reviewers score the same tools in the shared Excel workbook; answers combine by weight, then import here.
Brian
×1
weight
Mike
×2
weight
Allison
×2
weight
Van
×3
weight
1
Fill your tab
Open the Team Scoring workbook, go to your colored tab, and score each tool. Guests can use the Guest tab and set their own weight.
2
Combine
The Combined tab weight-averages everyone automatically (anyone who leaves a score blank is skipped).
3
Import
Click Import Team Excel in the toolbar and choose the workbook — it replaces the sample data with your team's weighted results.
§ Glossary
Enterprise-Grade
Software designed for large, regulated organizations.
FedRAMP
Federal authorization program for cloud security compliance.
Cybersecurity Posture
Overall security readiness of a system or vendor.
Data Governance
Policies and controls managing data quality, security, and usage.
PHI
Protected Health Information — sensitive health data requiring protection.
Deployment models
Cloud (vendor-hosted), On-Premises (local secure infra), Hybrid, or Edge.
Interoperability
Ability of systems to exchange and use shared data (e.g., IFC, COBie).
Digital Twin
Dynamic digital representation of a facility linked to performance data.
Generative Design
Algorithm-driven creation of optimized design options.
QA/QC
Quality Assurance / Control — processes ensuring compliance and accuracy (e.g., clash detection).
Computer Vision / NLP
AI that interprets images & video, or analyzes & generates human language.
Predictive Analytics
AI forecasting outcomes from historical data.
UFC Touchpoint
A UFC section or workflow potentially influenced by a tool.
Soft Threshold Flag
A non-disqualifying risk indicator in the framework.
Zero Trust
Security model requiring continuous verification of users and devices.
? Frequently asked
Where is my data stored?
In your team's shared cloud database (Supabase). Everyone signed in reads and writes the same data, and changes sync live between people who have it open. Opened offline as a local file, it falls back to this-device-only storage.
How do I add a new tool?
Click New Evaluation, fill the five sections, and Save. To do many at once, score them in the Team Excel workbook and import it.
Can I export results?
Yes — the Export menu gives a text summary, a single-tool CSV, an all-tools CSV, or a PDF (via print).
What does Reset do?
It clears the current form only. Your saved evaluations are not affected.
Why won't a secondary category select?
The secondary category can't be the same as the primary — choose a different one.
AI
Welcome to the AI Evaluation Matrix
Task #5 assesses existing AI products for the planning, design, construction, and operation of DoD medical facilities — judging what's feasible within DoD constraints and how each could inform UFC 4-510-01. New here? Take the 2-minute guided walkthrough, or explore on your own — full docs live under Help.
AI
Sign in
AI Evaluation Matrix · UFC 4-510-01
Accounts are created by your administrator. Contact them if you need access.
☁ Cloud Sync
Connect a shared Supabase database so your whole team reads and writes the same evaluations. Leave it disconnected to keep data only on this device.
Not connected — data is stored only on this device.