2025 GSA AI Use Case Inventory

Explore our AI use case inventory below, download the inventory [CSV - 28 KB], and view commercial off the shelf (COTS) [CSV - 2 KB] GSA AI use cases.

Use case name	What problem is the AI intended to solve?	Use case topic area	AI classification	Stage of development
Improve ASSIST Application Help	The AI is intended to solve the problem of users needing to sift through multiple documents to find help for specific ASSIST system functionalities. It aims to modernize application help by providing context-driven answers through an AI chatbot, leveraging existing help documents in various formats like PDF, Word, and Excel.	Information Technology	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
ASSIST Global Search and Inference	The AI is intended to solve the problem of inefficient information retrieval within the ASSIST system, where users currently spend up to 5 minutes per search due to a basic keyword-matching search solution. The goal is to reduce search time by 90% by implementing a context-aware, AI-powered search and structured data aggregation experience, allowing users to find accurate information using natural language in 30 seconds or less per query.	Information Technology	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
ASSIST Auto Copy and Summarize	The AI is intended to solve the problem of significant manual effort and time spent on data entry within the end-to-end acquisition workflow, where data is received in various file formats. Currently, data entry per contract typically takes between 45 minutes and one hour, and the proposed solution aims to automate this process to reduce per-contract data entry time by 80%.	Procurement & Financial Management	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
ASSIST ECF Autoclassification and Summary	The AI is intended to solve the problem of manual document review and categorization by Contracting Officers (COs) within the ASSIST Electronic Contract File (ECF) system. Currently, COs spend significant time reviewing and categorizing 20-50 documents per acquisition, leading to approximately 100-150 monthly help desk calls from new COs seeking guidance due to the manual effort and potential for errors.	Administrative Functions	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Leveraging Retrieval Augmented Generation (RAG) AI to Power Outcome-Based Contracting	The AI is intended to solve the problem of federal acquisition teams struggling to write clear, concise, outcome-based requirements, which often leads to incomplete or undesirable contract outcomes and higher costs. This experimental effort aims to determine if an AI-powered assistant can provide coaching and feedback on drafting such requirements with accuracy and reliability comparable to experienced human facilitators, thereby reducing reliance on limited Subject Matter Expert availability.	Procurement & Financial Management	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
SINSpector: Using Generative AI to Enforce Scope and Compliance	The AI, named SINSpector, is intended to solve the problem of efficiently and accurately identifying prohibited or out-of-scope items listed on GSA Advantage, as well as determining if products being added under a specific Special Item Number (SIN) are within the scope of that SIN. This addresses challenges such as inconsistent turnaround times, staffing reductions, and the manual effort involved in reviewing large volumes of products.	Procurement & Financial Management	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Content Management Analysis System Pilot	The AI is intended to solve the problem of automating data extraction from PDF, Excel, and Word documents stored in an isolated data lake. It aims to develop a prototype web application that can extract, structure, and enable focused analytics from IT procurement documents, and support plain language search and prompted interaction with the extracted data.	Procurement & Financial Management	Natural Language Processing (NLP)	b) Pilot – The use case has been deployed in a limited test or pilot capacity.
Acquisition Analytics	This AI is intended to solve the problem of classifying transactions within the Government-wide Category Management Taxonomy, which helps category managers accurately view the distribution of obligations and make decisions on aggregating multi-agency spend.	Procurement & Financial Management	Natural Language Processing (NLP)	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Category Taxonomy Refinement Using NLP	The AI is designed to address the challenge of maintaining an accurate, scalable, and contextually relevant taxonomy classification system for agency data. Traditional manual classification methods are time-consuming, prone to inconsistency, and difficult to scale across large datasets. By using Natural Language Processing (NLP), the AI automates and refines the categorization of content, ensuring that data is consistently and accurately classified according to evolving standards and organizational needs.	Other	Natural Language Processing (NLP)	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
FAS Catalog Platform - Virtual Assistant	The AI is intended to solve the problem of users struggling to quickly find help and support documentation within the extensive FCP system. Currently, users often resort to contacting the Vendor Support Center (VSC) instead of sifting through existing user guides, FAQs, and quick reference guides. The Virtual Assistant aims to streamline access to this vast volume of support information, providing quicker answers to common support questions and alleviating the burden on human support staff.	Service Delivery	Generative AI	b) Pilot – The use case has been deployed in a limited test or pilot capacity.
AI-Powered Visibility for Supply and Vendor Risk Resilience	The AI system is designed to solve the problem of managing risks in the federal supply chain and with third-party vendors by transforming traditional reactive risk management into a predictive and strategic capability. It addresses issues such as time-consuming manual evaluations, complexity in regulatory compliance (e.g., Buy American Act, CMMC), risks in federal procurement contracts (compliance, financial, geopolitical), procurement fraud, and disruptions from global events, ultimately aiming to ensure continuous access to essential supplies and protect against high-risk products or companies.	Procurement & Financial Management	Classical/Predictive Machine Learning	a) Pre-deployment – The use case is in a development or acquisition status.
AI in Public Experience (PX) Contact Center Services Blanket Purchase Agreement (BPA)	The AI is intended to solve the problem of high call, email, and chat volumes within the Public Experience (PX) Contact Center (CC) program, which currently handles over 2 million inquiries annually. By leveraging AI, the PX CC aims to decrease these volumes, thereby lowering operational costs for the Government. This aligns with the program’s mission as a “voice of the Government,” helping citizens navigate and understand Government programs, services, and information, especially as it is deemed a mission-essential function by GSA, expected to remain operational even during national crises.	Service Delivery	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
LaborMatch IQ - Efficient Services Pricing Market Research	The AI is intended to solve the problem of inefficient and ineffective services pricing market research. By leveraging a RAG (Retrieval-Augmented Generation) AI capability, the system aims to streamline and improve the accuracy of gathering market pricing information for services, which is a crucial aspect of procurement activities.	Procurement & Financial Management	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Login.gov: artificial intelligence technology used for detecting and mitigating fraud for remote unsupervised identity verification	The AI is intended to solve the problem of detecting and mitigating fraud, spoofing, or deepfakes introduced by the remote identity proofing process. Specifically, it aims to compare and verify a user’s identity by scanning identity documents uploaded online for authenticity, conducting one-to-one facial matching comparing the applicant’s self photograph (selfie) with the image on the identity evidence, performing liveness detection to mitigate imposters uploading images of stolen or counterfeit documents, and fraud detection analytics, all with the goal of ensuring the applicant is the person they are claiming to be. This is also to achieve NIST SP 800-63 compliance, which effectively requires AI-based capabilities for compliance with remote proofing.	Service Delivery	Computer Vision	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Enterprise Management Program Management Office Chatbot	The AI is intended to solve the problem of efficiently answering a wide range of compliance-related questions from across the organization, which currently consume significant team resources. These questions involve policy interpretations, statutory compliance obligations, internal compliance procedures, documentation requirements, and reporting guidelines. By automating responses to frequently asked questions and providing clear, actionable process steps, the AI aims to reduce the manual effort involved in question-answering and streamline the provision of compliance information.	Administrative Functions	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Public Comments Analysis	The AI is intended to solve the problem of labor-intensive and time-consuming processing of public comments on US agencies’ published notices, proposed rules, and deregulations, which are received in large amounts through OROS/eRulemaking managed regulations.gov. Agency users currently leverage FDMS.gov to process and respond to these comments, a task that is particularly challenging under short comment and response periods. The AI aims to streamline this process by providing analytical capabilities to increase efficiency and accuracy.	Administrative Functions	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
FAS Catalog Platform - Image Recognition for Product Photos	The AI is intended to solve the problem of poor quality or inaccurate product photos on the FAS Catalog Platform, which negatively impacts the buying experience for government purchasers and potentially sales for vendors. Currently, manual review of these images is impractical due to the large volume of products, leading to many substandard images remaining on GSA Advantage. By automating the assessment of image quality and accuracy, the AI aims to streamline the review process and improve the overall visual quality of product listings.	Procurement & Financial Management	Computer Vision	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Lot Description Generation	The AI is intended to solve the problem of simplifying sales creation and the lotting process by automatically generating lot descriptions. This will reduce sales contracting officer time, minimize manual errors, and help sell assets more quickly.	Administrative Functions	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Similar Auction Items Recommendations	The AI is intended to solve the problem of the current auction item search setup being time-consuming and involving repetitive efforts for users. It aims to enhance the auction user experience by offering added convenience through “Search Similar” or recommendations for items searched for auctions.	Administrative Functions	Classical/Predictive Machine Learning	a) Pre-deployment – The use case is in a development or acquisition status.
PPMS - Decode Image to Generate Property Description	The AI aims to simplify the sales creation and property reporting process for GSA by automating the drafting of GSAAuctions sale descriptions and filling in property details from images. This addresses challenges such as manual data entry time, human error in property reporting, and the impact of personnel losses.	Administrative Functions	Computer Vision	a) Pre-deployment – The use case is in a development or acquisition status.
ReDux Toolkit	The AI is intended to solve the challenges associated with application modernization and legacy system assessment, specifically by accelerating requirements generation automation, facilitating migration from current to target language and software stacks, and providing comprehensive documentation for all use cases and system functionalities, thereby addressing issues such as limited availability of Subject Matter Experts (SMEs).	Information Technology	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Gemini for Google Workspace	The AI (Gemini for Google Workspace) is intended to help federal agencies enhance productivity and streamline operations by assisting with tasks such as document drafting, data analysis, communication enhancement, and workflow automation. This production implementation aims to benefit users and support the agency’s mission to deliver value and savings in real estate, acquisition, technology, and other mission-support services.	Administrative Functions	Generative AI	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Enterprise Chatbot (GSAi)	The AI is intended to solve the problem of providing GSA staff with an enterprise-wide chatbot tool for general productivity support. This will allow users to access an internally available chatbot for tasks such as generating initial drafts, outlines, and copyediting for writers; facilitating faster research and more nuanced understanding of concepts for researchers; and drafting functions and working code for engineers.	Information Technology	Generative AI	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
USAi.gov	The AI is intended to solve the problem of making it easier for other agencies to rapidly deploy a chatbot to their organizations, as well as providing standardized API access to a suite of curated models and a mechanism for managing their user base. This addresses the challenge of individual agencies needing to build and maintain their own generative AI solutions, by offering a shared service model.	Information Technology	Generative AI	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Cybersecurity Chatbot	The AI is intended to solve the challenges associated with manual, labor-intensive compliance documentation and assessment processes within GSA IT. Currently, there’s a lack of effective LLM tools for compliance, and vendors are still developing solutions without definitive timelines. This leads to a persistent demand for compliance support that exceeds available supply and resources, hindering the ability to achieve the strongest possible security posture.	Cybersecurity	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
Elastic Machine Learning Threat Detection (Phase 1)	The AI is intended to solve the problem of identifying and detecting threats within system logs. By utilizing Elastic Machine Learning and Amazon SageMaker, the system performs time series analysis on log data to proactively identify potential security breaches or anomalies that might otherwise go unnoticed. This continuous analysis aims to enhance the overall security posture of GSA IT.	Cybersecurity	Classical/Predictive Machine Learning	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Elastic Machine Learning Threat Detection (Phase 2)	The AI is intended to solve the problem of effectively identifying and prioritizing cyber threats. By creating aggregation models and decision tree models, the AI aims to synthesize information from various sources, including Phase 1 models and new use cases, to provide a comprehensive understanding of user behavior and machine communication patterns. This will allow for the isolation and ranking of cyber threats, facilitating faster and more accurate escalation to threat hunting and incident response teams.	Cybersecurity	Classical/Predictive Machine Learning	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Slack AI	The AI is intended to solve the problem of inefficient team communication and productivity by providing AI-powered assistance within the Slack messaging platform. This includes addressing challenges such as managing long message threads, difficulty in finding relevant information, and optimizing workflows, ultimately aiming to reduce the time spent on message management and foster better knowledge sharing within teams.	Information Technology	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
FAS Vision Agentforce	The AI, specifically the “FAS Vision Agentforce” system, is intended to solve several key customer service challenges by automating and improving operations. Its primary purpose is to efficiently handle customer inquiries, especially routine ones like order status requests, and to automate discrepancy case creation for issues with orders received. By doing so, it aims to alleviate the burden on human customer service representatives, reduce manual processes, and provide quicker, more consistent support to agency customers.	Service Delivery	Agentic AI	a) Pre-deployment – The use case is in a development or acquisition status.
ServiceNow Generic Ticket Classification	Used to automatically route generic tickets to the correct group.	Information Technology	Classical/Predictive Machine Learning	a) Pre-deployment – The use case is in a development or acquisition status.
ServiceNow Virtual Agent (Curie)	This use case is designed for use within GSA to allow employees to research IT issues presenting to their workstation or find a quick answer to an IT procedure.	Service Delivery	Natural Language Processing (NLP)	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Marketplace: AI Supported Bookings & Space Stacking	The use case is intended to solve the problem of inefficient real estate resource management within federal agencies. It aims to optimize workspace utilization by providing smart recommendations for seating arrangements and space allocation, thereby reducing friction in the reservation process, minimizing unused space, and enabling data-informed decisions for space consolidation and cost reduction.	Administrative Functions	Classical/Predictive Machine Learning	a) Pre-deployment – The use case is in a development or acquisition status.
OCFO Chatbot	The AI is being developed to address the challenge of providing efficient and timely financial inquiry support by automating responses to common user questions. This aims to reduce the workload on financial personnel and improve the accessibility of accurate financial guidance, thereby streamlining the process and eliminating the need for users to wait for human intervention for routine inquiries.	Procurement & Financial Management	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
GovCXAnalyzer	The AI is intended to enhance the analysis of customer experience (CX) data across government services. It aims to process structured and unstructured customer feedback to identify key themes, sentiments, and pain points, thereby improving government services.	Service Delivery	Natural Language Processing (NLP)	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Strategic Atlas	The AI is intended to solve the problem of difficulty in monitoring strategy execution and related opportunities at GSA, specifically because execution often depends on agency partners whose priorities, progress, and results are hard to track.	Procurement & Financial Management	Generative AI	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
RPA Cloud Connection	The AI is intended to modernize the Process Optimization and Automation (PO&A) architecture by integrating the RPA Test environment with a cloud-based AI environment. This aims to automate the extraction and processing of data from unstructured documents such as invoices, receipts, forms, and emails, which typically consume significant manual effort and time. The goal is to streamline document-based workflows and improve overall efficiency in operations.	Information Technology	Computer Vision	a) Pre-deployment – The use case is in a development or acquisition status.
Office of Civil Rights (OCR) Case Crawler	The AI is intended to solve the problem of manually searching through extensive PDF documents of past EEOC OFO case decisions and OCR final agency decisions. This manual process is time-consuming and inefficient, and the AI aims to automate and streamline the search for desired terms or case names within these documents.	Administrative Functions	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
ChatOGC	The AI system aims to solve the problem of significant inefficiencies faced by attorneys within the Office of General Counsel (OGC) when searching for internal legal opinions and relevant GSA orders. These critical resources are currently scattered across various platforms, making them difficult and time-consuming to locate, leading to attorneys spending considerable time searching for precedents or contacting colleagues.	Administrative Functions	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
No-Code Text and Sentiment Analysis with XM Discover	The AI in this use case is intended to solve the problem of efficiently conducting text and sentiment analysis on qualitative feedback in a no-code environment. This allows GSA to gain a more nuanced view of customer sentiment and connect text feedback to key survey metrics and operational data for more thorough insights, moving beyond simple text summarization. Ultimately, the AI helps GSA to improve its ability to make data-driven decisions regarding customer needs.	Service Delivery	Natural Language Processing (NLP)	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Solicitation Review Tool (SRT)	The AI is intended to solve the problem of efficiently identifying non-compliant Information and Communications Technology (ICT) solicitations within SAM.gov data. It aims to automate the initial screening process to flag solicitations that may lack necessary compliance language, which would otherwise require extensive manual review by agencies.	Procurement & Financial Management	Natural Language Processing (NLP)	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Solicitation Review Tool - Section 508	The AI is intended to solve the problem of ensuring that federal solicitations for Information and Communication Technology (ICT) include adequate Section 508 compliance requirements to support accessibility in federal IT. It aims to automate the detection of ICT-related solicitations and then check for Section 508 compliance, improving the accessibility of government IT acquisitions.	Procurement & Financial Management	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
Software Supply Chain Security Research	The AI is intended to solve the problem of time-consuming manual research into security vulnerabilities for third-party software as part of the Software Security Review process. Currently, a manual process handles a limited number of issues per day, and a backlog of over 200 issues exists, with each issue potentially taking hours of human effort for research.	Cybersecurity	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
Expedited Transfer of Program of Requirements into Test Fit Layouts and Concept Design Options	The AI is intended to solve the time-consuming process of generating initial test fit layouts and concept design options from program requirements, a task that traditionally takes weeks to months. By leveraging AI, the system aims to significantly expedite this process, allowing for preliminary designs to be created within 24 to 72 hours. This addresses the need for faster turnaround times in initial design conversations and concept visualization for potential clients.	Service Delivery	Generative AI	a) Pre-deployment – The use case is in a development or acquisition status.
National Computerized Maintenance Management System (NCMMS) AI Chatbot	The AI is intended to solve the problem of providing user support within National Computerized Maintenance Management System (NCMMS) (IBM Maximo) for facility managers, contractors, and management teams, specifically by offering contextual support for the 20,000 work orders submitted daily.	Service Delivery	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
PBS AI Chatbot Domain Enhancements	The AI chatbot is intended to solve the problem of siloed information and time-consuming research across various Public Buildings Service (PBS) domains, including Portfolio Management, Facilities Management, Project Delivery, Acquisition, Client Strategy, Architecture & Engineering, and Strategy & Performance. By consolidating disparate data sources, policy documents, and internal guidance into a conversational, citation-rich assistant, the AI aims to streamline access to critical information and provide on-demand, data-driven insights for PBS personnel.	Service Delivery	Natural Language Processing (NLP)	a) Pre-deployment – The use case is in a development or acquisition status.
Leasing Desk Guide Bot	The AI, specifically the “Leasing Desk Guide Bot,” is intended to solve the problem of GSA stakeholders needing to efficiently access accurate, consistent, and context-relevant information from the Leasing Desk Guide to perform their duties. By leveraging a RAG-based internal chatbot, the system aims to streamline the information retrieval process, reducing the time and effort traditionally required to find specific details within a large corpus of knowledge, thereby enhancing operational efficiency for GSA personnel.	Administrative Functions	Generative AI	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Portfolio Analysis LLM	The AI is intended to solve the problem of enabling strategic decision-making around the optimization of the PBS real estate portfolio. This involves aggregating multiple data sources to answer key strategic questions, such as how to increase occupancy levels or find opportunities to decrease the cost per occupant within the portfolio.	Procurement & Financial Management	Generative AI	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Pricing Policy Chatbot	The AI, specifically the Pricing Policy Chatbot, is intended to solve the problem of efficiently providing factually accurate, context-relevant, and consistent information related to Pricing Policy to internal stakeholders. By training a RAG-based internal chatbot on a dedicated corpus of knowledge, the system aims to streamline the process of information retrieval and ensure that personnel have immediate access to the necessary guidance for their duties, thereby reducing potential inconsistencies and improving operational efficiency.	Administrative Functions	Natural Language Processing (NLP)	b) Pilot – The use case has been deployed in a limited test or pilot capacity.
Tier 1 Use Cases	These use case addresses challenges such as lack of familiarity with AI tools, inefficient low-stakes tasks, uninformed AI tool acquisition, and unaddressed ethical implications by providing a framework for exploring, evaluating, and understanding AI capabilities with non-sensitive information.	Administrative Functions	Other	c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.

2025 GSA AI Use Case Inventory

1 Choose a location

2 Choose a date