Educate, Eliminate Toil, Build Trust: How Federal Leaders Are Overcoming AI Hesitancy

How federal leaders sell AI internally - and what sales teams can copy

Federal agencies are learning how to beat AI hesitancy. Their playbook is simple: educate relentlessly, aim AI at work people dislike, and prove it works before scaling.

Skepticism isn't a bug. With five generations in the workforce, leaders expect it. As NASA's chief AI and data officer David Salvagnini said, there's good reason to be cautious and plenty of reason to be optimistic - both can be true.

The reality check: usage drives trust

AI users are still the minority in the U.S., and those who use it are about twice as likely to trust it compared with non-users, according to recent Gallup polling. Inside government, Department of Defense leader Mark Gorak said he often sees "almost no hands go up" when he asks who used AI that day - even as he pushes daily use.

Translation for sales leaders: adoption won't happen by memo. You need reps using AI in real work, daily, to build trust.

Lead with education - and make policy obvious

The State Department's internal chatbot, StateChat, didn't take off on day one. CIO Kelly Fletcher said the real bottleneck wasn't technical scale - it was getting people comfortable through focused training and constant communication.

Even with internal branding and a secure homepage, employees still asked if they were allowed to use it. Clear policies, fast-start guides, and hands-on demos changed that.

Sell the outcome people want: less toil

ARPA-H paired IT with HR and went team by team asking a simple question: what's the hardest, most time-consuming task? Then they pitched AI against that pain and backed it up with consistent messaging and open info sessions.

The Department of the Interior started with contract closeouts. No one fought it - because no one wanted to do that task. They now apply AI to grant closeouts and use tools such as ChatGPT and Microsoft Copilot Chat to help staff draft required summaries for public reporting on USAspending.gov. Management used AI first for oversight to set the bar and build credibility.

Build trust with framing and proof

NASA tested an LLM-based medical resource for astronauts, first named "Doc in a Box," then reframed as the "Crew Medical Officer Decision Assistant." Small change, big signal: this is augmentation, not replacement.

Doctors curated the data extensively to make outputs reliable. That process won over leadership and moved the tool from experiment to accepted resource.

The sales leader playbook

Borrow what works. Adapt it to your revenue team.

• Start where reps feel the pain: CRM hygiene, call notes, recap emails, proposal boilerplate, RFP compliance matrices, account research, and meeting summaries.
• Pick 2-3 use cases that save 30-60 minutes per seller per day. Prove time back before anything else.
• Make it safe to use: publish a plain-language policy, list approved tools, set data-handling rules, and define "no-go" content.
• Run weekly 45-minute live demos with real deals. Record snippets. Create 5-10 prompt templates your team can copy.
• Turn skeptics into advisors: invite them into pilots, ask for failure modes, and tighten controls based on their feedback.
• Brand the tool for context: "Deal Desk Assistant," "RFP Assistant," or "Call Recap Assistant." Names influence expectations.
• Measure and broadcast wins: time saved, faster proposals, improved forecast notes, lower zero-activity days. Share short before/after examples.
• Stand up a cross-functional squad: Sales Ops + RevOps + IT + Legal/Compliance for fast approvals and guardrails.
• Keep a human in the loop for anything customer-facing. Trust grows when quality stays high.

30-60-90 rollout

• Days 0-30: Inventory toil per role. Choose 1-2 safe tools. Publish policy. Pilot with 10-15 sellers. Create prompt packs for notes, emails, and research.
• Days 31-60: Expand to the team. Add office hours. Track time saved and usage. Share two wins per week in the sales channel.
• Days 61-90: Automate obvious steps (call recap to CRM, follow-up draft to email). Add light oversight checks for compliance and tone. Lock in what works.

What this means for you

Don't argue people into AI. Show them how to skip the work they dislike, make it safe, and prove results. The trust follows the usage.

If you want structured training and role-specific prompts, see our curated paths for different jobs here: Complete AI Training: Courses by Job.