Introduction
The General Services Administration (GSA) can leverage AI to significantly enhance its operational efficiency and service delivery. By implementing advanced AI technologies in areas such as procurement analytics, document processing, and customer support, the GSA can streamline its processes and improve accuracy across various functions. These initiatives have the potential to optimize government spending, refine classification systems, and enhance customer interactions. The following use cases illustrate how AI can be employed to advance the GSA’s capabilities and support its mission more effectively.
Use Cases
-
1. Acquisition Analytics
This project involves the analysis of detailed transaction data to classify each transaction according to the Government-wide Category Management Taxonomy. By organizing this data effectively, the initiative aims to enhance procurement processes and improve overall government spending efficiency.
-
2. City Pairs Program Ticket Forecast and Scenario Analysis Tools
This initiative utilizes segment-level data from the City Pair Program to generate near-term forecasts for air travel purchases, providing insights for the current and upcoming fiscal year. By analyzing data at various levels of granularity, including distinctions between Department of Defense and civilian travel, the program aims to enhance planning and budgeting for government travel.
-
3. Category Taxonomy Refinement Using NLP
This project employs natural language processing techniques to refine category taxonomy by extracting tokens from product descriptions. This process aims to more accurately define intended markets for Product Service Codes (PSCs), thereby improving the classification and management of government procurement activities.
-
4. Key KPI Forecasts for GWCM
This initiative focuses on generating near-term forecasts for key performance indicators (KPIs) by analyzing monthly historical data related to various components. The pilot program specifically targets total agency and category spending, with the potential for successful methodologies to be applied to additional KPIs in the future, enhancing overall financial planning and accountability.
-
5. Contract Acquisition Lifecycle Intelligence (CALI)
The Contract Acquisition Lifecycle Intelligence (CALI) tool is designed to enhance the efficiency of evaluating vendor proposals by utilizing automated machine learning techniques. When a Contracting Officer (CO) receives vendor proposals, they can easily upload the necessary solicitation documents to CALI, which will then analyze the proposals for compliance in critical areas such as format, forms, representations and certifications, and overall requirements. This streamlined process not only saves time but also ensures a more accurate evaluation, as designated team members can easily review and finalize results within the system. Currently, CALI is being trained using sample data from the Multiple Award Schedule (MAS) program to improve its effectiveness.
-
6. Chatbot for Federal Acquisition Community
The implementation of a chatbot within the GSA FAS National Customer Support Center (NCSC) aims to significantly enhance the customer experience by automating responses to frequently asked questions sourced from public knowledge articles. This innovation is expected to reduce the need for live chat staffing, allowing NCSC personnel to focus on more proactive customer service initiatives. Importantly, customers will still have the option to connect with a live agent if they prefer personalized assistance, ensuring that all user needs are met effectively.
-
7. Document Workflow / Intelligent Data Capture and Extraction
The GSA is working to develop a more accurate and scalable document workflow platform that intelligently captures and classifies essential data from both structured and unstructured documents, particularly PDF files. This initiative aims to streamline the transfer of critical information to the appropriate processes, workflows, or decision-making engines, thereby enhancing operational efficiency and data management across the organization.
-
8. Service Desk Generic Ticket Classification
This project focuses on developing a classification model for generic Service Desk tickets to automate the re-routing process to the appropriate teams. Currently, this task is performed manually, which can be time-consuming and inefficient. By targeting the five most common ticket types, the model aims to streamline operations, reduce response times, and improve overall service efficiency within the organization.
-
9. Service Desk Virtual Agent (Curie)
The Service Desk Virtual Agent, named Curie, is a natural language chatbot designed to enhance the customer service experience for IT service requests. Utilizing machine learning, Curie provides predictive responses based on user chat entries, effectively leveraging a repository of knowledge-based articles. This virtual assistant aims to improve response accuracy and efficiency, making it easier for employees to access the information they need for their IT inquiries.
-
10. Solicitation Review Tool (SRT)
The Solicitation Review Tool (SRT) processes data from SAM.gov related to Information and Communications Technology (ICT) solicitations, compiling it into a database for analysis by machine learning algorithms. The initial application of this system involves a Natural Language Processing model that assesses whether solicitations include necessary compliance language, marking those that do not as non-compliant. Agencies are encouraged to review their data and validate the predictions made by the SRT, which is further supported by monthly random manual reviews conducted by GSA to ensure accuracy and reliability.
-
11. Classifying Qualitative Data
USAGov and USAGov en Español gather extensive qualitative data from various sources, including survey comments, web searches, and call center chat transcripts. This project focuses on categorizing these comments by topic, which helps identify areas that require product updates or enhancements. By analyzing this qualitative feedback, the organization can better understand user needs and improve service offerings accordingly.
-
12. IAE FSD CCAI Virtual Agent
This Virtual Agent, named SAM, is designed to understand and respond to customer inquiries effectively through manual learning techniques. By utilizing natural language processing, SAM aims to provide accurate and relevant responses to customer needs, enhancing the overall user experience and support provided by the agency.