Learning Objectives

After completing this lecture, you will be able to:

• Understand the key concepts of change management for ai adoption in a government context
• Connect change management for ai adoption to your agency's AI initiatives
• Identify next steps for applying these concepts in your role

Key Topics Covered

-
Overcoming resistance

-
Training strategies

-
Champions and early adopters

-
Addressing fear and uncertainty

Why This Matters for Government

Government agencies face unique challenges when it comes to AI adoption. This lecture addresses these challenges head-on by providing analysts, project leads, team supervisors with the knowledge and frameworks needed to navigate AI in the public sector responsibly and effectively.

As part of the L2 (AI Practitioner) curriculum, this lecture builds on the foundational principle that every AI system in government ultimately serves citizens. Whether you are working with AI tools daily or setting strategy for your agency, understanding change management for ai adoption is essential for responsible, effective government AI adoption.

======================================================================

TRANSCRIPT: Change Management for AI Adoption

======================================================================

Chapter: 4 -- AI Project Management

What you will learn:

• Why AI adoption requires change management, not just technology deployment
• Identifying and overcoming resistance to AI adoption
• Building AI champions within your organization
• Training strategies for different audience levels
• Phased rollout approaches that manage risk and build confidence
• Measuring adoption and engagement
• Creating organizational readiness before deploying systems

You've built a great AI system. It's accurate, fair, validated. Now you deploy it and discover that nobody's using it. Or they're using it but overriding it on every decision. Or they're using it but not understanding what it does, so they're making bad decisions based on it.

This is where change management comes in. Deploying technology is one thing. Getting an organization to actually use it well is something else entirely. This lecture teaches the change management practices that ensure AI systems get adopted and used effectively.

PURPOSE STATEMENT

Change management ensures that deployed AI systems are actually used and used well. Without change management, you get underutilization, misuse, or rejection. With good change management, you get organizational adoption that delivers the business value the system was designed to provide.

WHY THIS MATTERS FOR GOVERNMENT

Government agencies are conservative about change. Staff have been doing their jobs a particular way for years. AI systems ask them to change how they work. This creates natural resistance. Without addressing that resistance, systems sit unused. With good change management, teams adapt and the organization benefits.

Additionally, misuse is a serious risk. If staff don't understand how an AI system works, they might apply it in ways it wasn't designed for or trust it too much. Change management includes training and support that prevent misuse.

UNDERSTANDING RESISTANCE TO AI

Overview

Resistance to AI adoption is normal and predictable. Understanding it helps you address it effectively.

SOURCES OF RESISTANCE

Fear of job loss: "This AI will replace me"

• Address through: Clear communication that AI is augmentation not replacement, redeployment planning, skills training
• Reality: AI typically changes jobs, not eliminates them. But staff worry this anyway.

Loss of autonomy: "I'll have to follow what the AI recommends"

• Address through: Emphasizing human override (staff can always overrule the system), positioning AI as support not authority
• Reality: Staff should override the system when they have good reasons. Help them understand when override is appropriate.

Skepticism about fairness: "This AI is biased"

• Address through: Showing fairness analysis and validation results, addressing concerns transparently
• Reality: Some skepticism is healthy. Show evidence that you've tested for bias.

Complexity: "I don't understand how this system works"

• Address through: Clear explanations of what the system does (not HOW it does it), training, documentation
• Reality: Staff don't need to understand neural networks. They need to understand what inputs matter and what output means.

Workflow disruption: "This changes how I do my job"

• Address through: Designing systems that fit workflows, phased introduction, involving staff in design
• Reality: Change is uncomfortable. Support people through it.

ADDRESSING RESISTANCE

Don't ignore it. Resistance contains information about real problems that need solving. When staff say "This won't work because of X," listen. They might be right.

Engage early. Include staff in design and testing decisions. They become invested and better understand the system.

Communicate clearly. Explain what the system is and isn't, how it will change their work, and what's expected of them.

Provide support. Training, documentation, help desk support, peer mentors--whatever people need to adopt the system.

Show respect. Acknowledge that change is hard. Don't dismiss concerns as technophobia.

BUILDING AI CHAMPIONS

Overview

Champions are staff members who understand the system, believe in it, and help others adopt it. They're critical to successful change.

WHO MAKES A GOOD CHAMPION

Not necessarily the technically expert person. Champions need to:

• Understand the business problem the system solves
• Be respected by their peers
• Be willing to help others
• Be honest about system limitations
• Be comfortable with technology but not necessarily an expert

DEVELOPING CHAMPIONS

• Identify potential champions early (before system is deployed)
• Involve them in testing and validation
• Give them deep training and access to experts
• Make them part of the deployment team
• Give them authority to help others
• Publicly recognize their role and expertise

CHAMPION NETWORK

For large deployments, build a network of champions:

• Each office/team has a champion
• Champions meet monthly for training and sharing
• Champions help with local training
• Champions are first line of support for their peers
• Champions feed back problems to the central team

Champion support:

• Provide champions with detailed documentation
• Give them training on how to teach others
• Support them with escalation paths for difficult questions
• Acknowledge their time and effort
• Make them feel part of the deployment team

TRAINING AND CAPABILITY BUILDING

Overview

Training for AI systems is different from training for traditional software because people need to understand both how to use the system and what it does/doesn't do.

TRAINING AUDIENCES

Executive/Leadership:

• Business case for the system
• What problems it solves
• What success looks like
• How to monitor progress
• 30-60 minute training

Managers/Supervisors:

• How the system changes workflows
• How to manage staff adoption
• How to handle overrides and exceptions
• Performance metrics to monitor
• What to do if the system isn't working
• 2-3 hour training

End users (staff who use the system):

• What the system does in their specific context
• How to use it step-by-step
• When to trust it and when to override it
• What to do with unusual cases
• Where to get help
• 4-8 hours training (hands-on)

Auditors/Compliance:

• System documentation and validation
• How to audit it
• What evidence exists that it's fair and accurate
• 2-4 hours

QA/MONITORING STAFF

• What metrics to monitor
• What constitutes a problem
• How to investigate issues
• 4-8 hours training

TRAINING FORMAT

Classroom training: Good for building community, interactive learning, Q&A

Online training: Good for reach, asynchronous access, self-paced

Hands-on practice: Essential--people need to actually use the system with guidance

Job aids/documentation: Critical--people forget training and need references

Peer mentoring: Most effective for actual behavior change

Simulation/sandbox: Great for low-risk practice before live deployment

PHASED ROLLOUT AND PILOT APPROACHES

Overview

Don't deploy to the entire organization at once. Phased rollout manages risk and builds confidence.

PHASED ROLLOUT APPROACH

Phase 1: Pilot (weeks 1-4)

• Deploy to 1-2 offices or 2-3 champion users
• Very close monitoring
• Daily check-ins
• Quick adjustments if needed
• Learning focus: Does the system work? Do people use it?

Phase 2: Early adoption (weeks 5-12)

• Expand to 3-5 offices
• Weekly check-ins
• More independent operation
• Learning focus: Does it scale? What training do people need?

Phase 3: Broader rollout (weeks 13-20)

• Expand to 50% of organization
• Biweekly check-ins
• Learning focus: How are adoption rates?

Phase 4: Full rollout (weeks 21+)

• Expand to entire organization
• Monthly check-ins
• Ongoing support and continuous improvement

DECISIONS AT EACH PHASE

At the end of each phase, decision points:

• Are we proceeding to next phase? (Yes, yes with conditions, no/redesign)
• What did we learn? (What changes are needed?)
• What support do users need? (Training, documentation, tools?)
• What system changes are needed? (Performance, usability, features?)

Don't move to next phase if pilot has issues. Fix problems before expanding.

MEASURING ADOPTION AND ENGAGEMENT

Overview

You need metrics for adoption and engagement, not just system performance.

ADOPTION METRICS

Usage rate: What percentage of staff are using the system? (Target: 80%+ within 2 months)

Frequency: How often are people using it? (Daily, weekly, monthly?)

Coverage: What percentage of applicable cases are going through the system? (Is staff bypassing it?)

Consistency: Are people using it the same way across offices? (Or is there wide variation?)

ENGAGEMENT METRICS

Confidence level: Do people trust the system? (Survey staff)

Understanding: Do people understand what the system does? (Test their knowledge)

Adoption pace: Are we progressing through phases on schedule?

Training completion: What percentage of staff completed training?

OPERATIONAL METRICS

Override rate: How often do people override system decisions? (Some overrides are good, too many is concerning)

Error rate: What percentage of system outputs are wrong? (Catching system problems)

Time savings: How much time is the system actually saving? (Did it deliver promised value?)

Error resolution: How quickly are problems fixed?

METRICS DASHBOARD FOR LEADERSHIP

Share monthly:

• Usage rate by office/team
• Overall accuracy and fairness metrics
• Issue tracking (problems reported, resolved, outstanding)
• Training completion
• Adoption trajectory (on track?)

This keeps leadership informed and enables early detection of adoption problems.

ANTI-PATTERNS

ANTI-PATTERN 1

System is technically ready so you deploy it. You assume if it's good, people will use it. Nobody's using it because nobody knows about it or understands it.

How to avoid it: Build change management into project plan from the beginning. Allocate resources and time for adoption.

ANTI-PATTERN 2

You give people a 2-hour training and expect adoption. Adoption requires ongoing support, reinforcement, and community building.

How to avoid it: Change management includes training, but also communication, champions, support structures, and cultural change.

ANTI-PATTERN 3

Central team designs change, tells offices what's happening. Offices don't feel heard and resist.

How to avoid it: Involve local leaders and champions. Adapt change approach to local context. Listen to feedback.

ANTI-PATTERN 4

You deploy to the entire organization simultaneously. Problems emerge across 50 offices at once. You can't respond.

How to avoid it: Phase the rollout. Learn from each phase before expanding.

ANTI-PATTERN 5

Some people resist adoption. You ignore them and try to work around them. They quietly continue old processes.

How to avoid it: Engage resistant staff. Understand their concerns. Give them space to adapt at their own pace (within reason).

PRACTICE PROMPTS

EXERCISE 1

Design a change management plan for deploying an AI benefits eligibility system across 50 field offices:

• How would you manage resistance?
• Who would you engage as champions?
• What training would you provide?
• How would you structure the phased rollout?
• What metrics would you track?

EXERCISE 2

How would you specifically address this concern: "This system will let us deny benefits to people who deserve them because the AI is wrong"?

EXERCISE 3

You want to build a champion network for adoption. How would you:

• Identify champions
• Develop them
• Support them
• Connect them
• Measure their effectiveness

KEY TAKEAWAYS

• Change management is as important as technology for adoption success. Deploy technology without change management and it sits unused.
• Resistance is normal and contains useful information. Listen to it and address legitimate concerns.
• Champions accelerate adoption. Invest in identifying and developing champions who can help peers.
• Training alone isn't sufficient. Combine training with job aids, peer support, champions, and ongoing help.
• Phased rollout manages risk and builds confidence. Don't deploy to the entire organization day one.
• Measure adoption, not just system performance. Track usage, engagement, and adoption pace.
• Adapt to local context. What works in one office might not work in another. Involve local leaders.
• Change takes time. Don't expect full adoption in week one. Build in 3-6 months for organizational adoption.

GLOSSARY

Change Management: Systematic approach to managing organizational transition from current state to desired future state, including communication, training, and cultural change.

Adoption Rate: Percentage of target population actually using the system as intended, measured over time.

Champion: Staff member who understands the system deeply, is respected by peers, and helps others learn and adopt the system.

Resistance: Opposition or reluctance to change, stemming from various sources such as fear, skepticism, or workflow disruption.

Phased Rollout: Gradual deployment of a system across an organization in planned stages, starting with pilots and expanding based on learning.

Change management transforms technology deployment from a one-time event into an organizational journey. When you do it well, systems get adopted and deliver value. When you skip it, systems sit on shelves.

The most effective approach treats adoption as a shared responsibility: central team provides strategy and support, local leaders and champions drive adoption in their context, staff get training and help, and leadership stays engaged and responsive.

Reflect on an organizational change you've been part of (new process, new system, new policy). What helped adoption? What hindered it? How would better change management have improved the outcome? Use this reflection to build intuition for managing AI adoption.

You've learned how to manage organizational change for AI adoption. The next lecture focuses on measuring AI impact--how to know whether deployed systems are actually delivering the value they promised. See you there.

Government AI CLUB Certification Program

Level 2: AI Ready | Change Management for AI Adoption | Lecture 2.4.6

A GOVT.CLUB initiative

Visit: https://govt.club/learn/lectures/l2/246-change-management.html

======================================================================

<- 2.4.4 Testing and Validating AI Systems
2.4.6 Measuring AI Impact ->

Start Your CLUB Certification

This lecture is part of L2: AI Practitioner -- 40 hours of comprehensive government AI training.

Explore CLUB Certification

Related Lectures

L2
2.4.1 -- How AI Projects Differ from Traditional IT
60 min - Video + Comparison

L2
2.4.2 -- Requirements Gathering for AI
60 min - Workshop

L2
2.4.3 -- Working with AI Vendors and Contractors
60 min - Video + Checklist