Operational Excellence (OPEX) Insight – Thursday - April 23, 2026: Lost 40% of Its People, GSA Hands 1 Million Work Hours to AI.
Góc Nhìn Vận Hành Xuất Sắc – Thứ Năm, Ngày 23/04/2026: Mất 40% Người, GSA Giao 1 Triệu Giờ Việc Cho AI.
Welcome To Operational Excellence (OPEX) Insight Article For The Paid Subscriber-Only Edition.
This is the bilingual post in English and Vietnamese. Vietnamese is below.
Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.
English
PART 1 – OFFICIAL INFORMATION
On April 14, 2026, at the OpenText Government Summit in Washington D.C., General Services Administration (GSA) Deputy Director Michael Lynch officially announced the “Million Hours Challenge” — a program targeting the use of an internal artificial intelligence tool called USAi to automate 1 million work hours performed by federal employees and contractors. This declaration came against the backdrop of GSA having lost nearly 40% of its total workforce since October 2024, transforming the largest real estate and procurement management agency of the US federal government into an organization operating with only approximately 60% of its prior staffing levels. Lynch disclosed that the program had identified approximately 400,000 hours eligible for automation — nearly half the target — and that if successful, the model would be expanded across the entire federal government, making GSA the first national-scale AI operations laboratory in the history of American public administration.
To understand the scale of the workforce crisis that led to the Million Hours Challenge, one must look back at the chain of events beginning in early 2025. Under directives from the Trump administration and the Department of Government Efficiency (DOGE) led by Elon Musk, GSA became one of the most aggressively downsized federal agencies. Before the reductions, GSA had approximately 12,000 to 12,939 employees (per March 2024 data). The contraction occurred in multiple phases and forms: voluntary early retirement programs, hiring freezes, Reduction in Force (RIF — structured layoffs), a mandate requiring all employees to return to the office by March 3, 2025 (completely ending remote work), and consolidation of all regional offices into four to five centralized “hubs.” The Public Buildings Service (PBS) division — the unit managing the largest federal real estate portfolio in the United States — shrank from over 5,600 employees to approximately 3,100 during 2025, losing 45% of its workforce in roughly the period from September 2024 to November 2025. The Technology Transformation Services (TTS) division — the unit responsible for technology modernization across the federal government — was cut by 67% since January 2025. The renowned technology consultancy team 18F, once considered the symbol of digital innovation in the public sector, was completely dissolved in March 2025 when approximately 70 researchers, product managers, and specialists received termination notices early on a Saturday morning.
A Government Accountability Office (GAO) report published in April 2026 painted a deeply concerning picture of the operational consequences of these cuts. The GAO found that GSA conducted staff reductions before analyzing what skills it needed to fulfill its mission — in other words, cut first, assess later. The report identified four critical operational failures: GSA leadership did not develop performance measures to gauge the success of the overhaul; communication with stakeholders and employees about changes was conducted after changes were implemented rather than before; the agency lacked a systematic plan to monitor reform implementation; and analysis of how many personnel with specific skills were needed to fill gaps was performed only after cuts had already taken place. Real-world consequences cascaded to other federal agencies that depend on GSA services: one agency had to extend project timelines because GSA no longer had sufficient staff to conduct cost estimates before a building could be transferred; a property sale stalled because the team managing the transaction had left GSA, remaining staff were unclear whether steps to relocate a tenant agency had been completed, and GSA was forced to restart the entire disposal process from scratch.
It is within this context that the Million Hours Challenge emerged not merely as a technology initiative but as an operational survival effort. The strategy Lynch described follows the EOA framework — Eliminate, Optimize, Automate: first eliminate unnecessary processes, then optimize remaining ones, and finally automate what can be handled by AI. The USAi tool — GSA’s internal AI platform — serves as the centerpiece of this effort, targeting “repetitive, manual workflows” to free remaining employees to focus on serving customer agencies, improving procurement outcomes, and delivering mission-critical services. Lynch reported that internal response was “amazing,” with over 300 employees applying to participate, narrowed to an initial cohort of 30 staff addressing five core operational problem statements with a goal of creating “interoperable systems and processes.” Simultaneously, GSA also began rehiring approximately 400 employees over the next six months in facilities management, acquisition, and project management — a tacit acknowledgment that the cuts went too far.
