What happened
OPM released FWD Chat, a beta AI chatbot that allows users to query federal workforce data using plain English. The tool, built on OpenAI's GPT-5.6, translates user prompts into SQL queries to retrieve data from the agency's Enterprise Human Resources Integration datasets. It is currently in a testing phase where officials are soliciting public feedback to improve accuracy and usability.
The Office of Personnel Management (OPM) officially released FWD Chat, or Federal Workforce Data Chat, to the public on Wednesday in a beta testing phase. The tool is designed to answer questions about federal workforce data in plain English, such as hiring numbers, average salaries by state, and employee separation types. Adam Starr, OPM's chief information officer, announced the tool on LinkedIn, encouraging users to 'try to break it' and provide feedback.
According to Starr, the chatbot is built on OpenAI's GPT-5.6 model and uses a data programming language, specifically SQL, to interact with the agency's data. The project took approximately three months to complete. When a user asks a question, the AI translates it into a SQL query, runs it against the agency's Enterprise Human Resources Integration (EHRI) datasets, and returns the result. Each response includes a 'How was this generated' dropdown that displays the exact SQL query used, providing transparency into the AI's reasoning process.
The tool is currently limited to data available through the Federal Workforce Data website, which was overhauled in January to replace the legacy FedScope tool. It does not yet include data from interactive charts that are not published as record-level data, such as administrative leave use or federal recruitment details. Starr noted that the chatbot cannot yet combine data from different tables, a limitation the agency hopes to address in future updates.
OPM plans to collect feedback over the next month to iterate on the tool, with updates released twice a month. The agency intends to make the Model Context Protocol (MCP) server publicly available and coordinate with the America.gov team to integrate FWD Chat into the administration's broader federal services directory. This move aligns with a government-wide effort to use AI to provide authoritative and timely results for public inquiries.
Source details: fedscoop.com ↗
Why it matters
This launch represents a significant shift in how federal data is accessed, moving from static dashboards to interactive, AI-driven . By exposing the underlying SQL queries, OPM promotes transparency and allows users to verify the logic behind AI-generated answers. The tool is part of a broader effort to integrate AI into federal services, with plans to connect it to the America.gov platform and make its Model Context Protocol server public, potentially setting a standard for other agencies.
The launch of FWD Chat marks a practical application of large language models in federal data accessibility. By converting natural language into SQL, OPM lowers the technical barrier for citizens, journalists, and researchers to access workforce statistics. The transparency , which reveals the underlying code, addresses common concerns about AI hallucinations by allowing users to audit the data process.
This initiative is part of a broader trend of federal agencies adopting AI to modernize public-facing services. The integration with America.gov suggests a strategic move to centralize federal information access through AI interfaces. The public release of the MCP server could influence how other agencies structure their data for AI consumption, potentially standardizing AI-data interactions across the government.
However, the tool's limitations highlight ongoing challenges in federal data governance. The exclusion of certain data categories, such as race and ethnicity data removed from previous systems, and the inability to cross-reference different data tables, indicate that AI tools are only as comprehensive as the underlying data infrastructure. The beta phase is crucial for identifying these gaps and refining the model's ability to handle complex or nuanced queries.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
What to watch next
Monitor the next month of beta feedback to see how OPM addresses error rates and unanswered questions. Watch for the integration of FWD Chat with America.gov and the public release of the MCP server. Additionally, observe whether OPM expands the data sources to include previously excluded categories like race and ethnicity data or administrative leave statistics.
The next month of beta testing will be critical for assessing the tool's reliability. OPM is specifically monitoring error messages and unanswered questions to identify areas for improvement. The frequency of updates, planned for twice a month, suggests a rapid iteration cycle that may quickly address user-reported issues.
The integration of FWD Chat with America.gov is a key development to watch. This integration could significantly expand the tool's reach and utility, making federal workforce data accessible through a single, unified AI interface. The public availability of the MCP server will also be important for developers and other agencies looking to implement similar AI-data solutions.
Policy experts are watching to see if OPM addresses the absence of certain data categories, such as race and ethnicity data, in future updates. The tool's ability to answer questions about specific workforce actions, such as probationary terminations, will also be a point of scrutiny, particularly given the current administration's workforce policies.