The Daily Scoop Podcast: Inside GSA’s Rollout of USAi

Date: September 30, 2025
Host: Billy Mitchell
Guests: Zach Whitman (Chief Data Scientist and Chief AI Officer, GSA), Nadia Hanson (Global Digital Transformation Executive, Salesforce)

Episode Overview

This episode spotlights the General Services Administration’s (GSA) launch of USAi, a new shared platform to streamline federal access to commercial AI models. Host Billy Mitchell interviews GSA’s Chief AI Officer Zach Whitman for an exclusive discussion on USAi’s goals, architecture, key challenges, and scaling strategy. A sponsored segment then features Salesforce’s Nadia Hanson on how “agentic AI” is reshaping public sector teams, emphasizing the importance of governance, workforce training, and strategic AI adoption.

Key Discussion Points

1. GSA’s Role in Government AI Adoption

[05:32] – [07:10]

• Enabler, Not Just User:
  Zach Whitman highlights GSA’s mission as "an enabler" for federal agencies, helping them "bootstrap into safe, transparent, generative AI practices” (05:37).
• Specialization & Scale:
  Many agencies lack resources or specialized knowledge to implement AI securely; GSA provides centralized, scalable solutions so agencies can focus on their unique missions rather than building AI infrastructure from scratch.

Quote:

“It’s unfair to put all that on the other agencies to take on from ground zero... Our job is to enable the safe adoption of these tools and make it so that CIOs, CISOs...have an easy option for a safe approach to empower their workforce.” — Zach Whitman (06:17)

2. USAi Platform Highlights – Features, Benefits, Challenges

[07:10] – [09:39]

• Marketplace & Accessibility:
  USAi is structured as a marketplace grounded in GSA’s award schedule acquisition platform.
• Low-Cost, Low-Barrier Access:
  Agencies can freely try tools, lowering the upfront cost and risk of AI experimentation.
• Vendor & Cloud-Agnostic:
  USAi prevents “walled garden” IT scenarios—no matter which cloud or environment agencies use, they can access leading AI models on a “very fair playing field” (07:56).
• Safety and Fit:
  Emphasis on constraining models for safety and agency-specific purposes.

Quote:

“We wanted to make sure that it was a very fair playing field and the models spoke for themselves...we just provide that conduit for safety.” — Zach Whitman (07:50)

Challenges/Drawbacks:

• Less control for agencies:
  Shared service means agencies have limited autonomy/flexibility, which may not accommodate highly niche applications (08:54).
• General-purpose focus:
  Platform targets broad use, not every specialized need—some requests/resources might not be prioritized.

3. Collaboration and Leveraging Agency Innovation

[09:39] – [11:08]

• Partnerships:
  GSA is collaborating with NIST, CISA, and federal scientific agencies to pool knowledge and accelerate best practices in model evaluation and deployment.
• Sharing Homegrown Innovation:
  GSA aims to make agency-developed models more widely available (e.g., Argonne National Lab’s LLM for scientists).

Quote:

“There’s a ton of really cool work being done across the federal complex. ...We think that this is an opportunity for us to use our scale to the broader good.” — Zach Whitman (10:37)

4. Addressing Scale and Ensuring Success

[11:08] – [12:41]

• Built for Internal Use, Scaled Out:
  The initial impetus was to serve GSA’s own culture; expanding to USAi was a logical next step.
• Market-Driven Future:
  GSA doesn’t see itself as the long-term provider but as a bridge until the commercial market matures.
• Rapid Model Turnover:
  The platform is designed for quick acquisition/deployment of new AI models, minimizing lag in access for agencies.

Quote:

“We wanted to make sure that the agencies didn’t have any kind of acquisition barriers or technical challenges that would impede their workforce from using the best class.” — Zach Whitman (12:21)

5. Keeping USAi Current with AI Model Evolution

[13:02] – [13:44]

• Pipeline for New Models:
  Process includes safety and performance evaluations.
• Empowering Agencies:
  Each “tenant” agency of USAi can run their own tests to verify fit.
• Automated, Accelerated Adoption:
  Minimizing “time to market” is a priority; steps are being automated to quickly onboard emerging AI models.

Quote:

“Our main mission is to minimize the time to market on model availability. Once it’s released, we want to bring it in as quickly as possible.” — Zach Whitman (13:33)

Salesforce Segment: Agentic AI in Government (with Nadia Hanson)

1. Workforce Adaptation Challenges

[15:08] – [17:18]

• Capacity Constraints:
  Governments must “do more with less”; AI’s greatest promise is supplementing limited human resources.
• Importance of Training:
  Technology shifts are about people, too. Programs like Salesforce Trailhead “democratize” skills training so any staff member can build AI literacy.

Quote:

“Teams need clarity on what skills they should develop first and reassurance that AI is here to augment rather than replace...” — Nadia Hanson (16:27)

2. Governing “Agentic AI”

[17:18] – [20:04]

• From Prediction to Autonomy:
  Traditional AI predicted outcomes; agentic AI can now act autonomously—leaders must plan for oversight and trust from the outset.
• Playbook for Adoption:
  • Top leadership sponsorship
  • Cross-functional oversight committees
  • Clear, small-scale first use-cases
  • Human oversight (“humans in the loop”)
  • Measuring and iterating after trust/impact are proven

Quote:

“Agent AI is different. These are autonomous agents that can take action. They don’t just recommend, they interact like a human executing tasks. So that's a huge opportunity. But it also means leaders must be very intentional about trust, about oversight, about governance from the very beginning.” — Nadia Hanson (17:38)

3. Practical Use Cases for Agentic AI

[20:04] – [21:36]

• Social Services Casework:
  AI agents can assemble summaries from disparate sources, freeing caseworkers to focus on people.
• Permitting and Licensing:
  AI handles intake, flags missing documents, drafts approval notes—letting staff handle complex reviews.
• Emergency Management:
  AI triages urgent calls to first responders faster than manual processes.

Quote:

“In each of these examples, the worker always stays in control, with AI being a digital helper, not a replacement.” — Nadia Hanson (21:18)

4. Foundation for Effective AI Adoption

[21:36] – [23:10]

• Start with Skills and the Right Use Case:
  Prioritize digital/data literacy and begin with high-value, low-risk tasks.
• Establish Governance:
  Data, ethics, and accountability policies are critical.
• Human Final Oversight:
  Employees must always have the final say, ensuring AI augments, not replaces, human judgment.

5. The Future of Government Work

[23:10] – [25:10]

• Administrative Tasks Automated:
  AI will shift government workers’ time from routine tasks to strategic, problem-solving and human-centered services.
• Changing Roles:
  New categories of “forward deployed” engineers and cross-disciplinary roles blending tech fluency with mission delivery will emerge even in public service.

Quote:

“It’s also going to redefine what talent looks like in government. ... More people who can think critically about data ... and that honestly is a big cultural shift.” — Nadia Hanson (24:02)

Notable Quotes & Memorable Moments

• On value of shared services:
  “It’s not central to their mission...our job is to enable the safe adoption of these tools.” — Zach Whitman (06:14)
• On agency collaboration:
  “We think that this is an opportunity for us to use our scale to the broader good.” — Zach Whitman (10:37)
• On new work paradigms:
  “AI will shift the focus ... to being more strategic ... an opportunity to make government jobs more meaningful and impactful.” — Nadia Hanson (23:24)

Timestamps for Key Segments

• [04:29] – [13:44]: Zach Whitman interview on GSA/USAi
• [14:19] – [25:29]: Agentic AI segment with Nadia Hanson

Conclusion

This episode offers a nuanced look at how GSA is centralizing and simplifying AI access for federal agencies via USAi, addressing both efficiency and safety. It also explores broader implications of autonomous AI in government and the importance of treating AI adoption as a strategic, human-centered, and well-governed process. Both segments underscore the need for continual learning, collaboration, and scalable yet flexible approaches as AI fundamentally transforms government work.