Episode 2510 New

How to Script Change Before It Derails

The concept of a "change script"—a practical tool for guiding people through transitions, especially when tech like AI is involved

Episode Summary

In this episode of Accelerating Humans, Julianna and Bert break down the concept of a "change script"—a practical tool for guiding people through transitions, especially when tech like AI is involved. Drawing inspiration from an MIT case study, they explore why even the smartest strategies can flop without a clear narrative.

They share real-life lessons from AI rollouts and corporate transformations, and offer a simple framework leaders can use to get buy-in, ease confusion, and avoid resistance. From the "could vs. should" framework to understanding when to use AI versus automation, this episode cuts through the hype with practical guidance.

Whether you're leading change, reacting to it, or just trying to make sense of what's next, this episode offers sharp insights—and a few hard truths—about what it really takes to move people forward.

Episode Timeline

0:00

Welcome & Course Announcement

Julianna opens with gratitude for listeners and reminds everyone about the Automate to Accelerate presale at acceleratinghumans.com/presale.

0:40

Introducing the Change Script

What is a change script? Why rollouts fail when people are confused, resistant, or checked out. The MIT-backed concept that gives teams a narrative, not just a mission statement.

2:54

The MIT Article: 95% Failure Rate

Julianna discusses the Fortune article about MIT's finding that 95% of generative AI pilots at companies are failing. Bert questions: how are they screwing this up?

4:51

Change Management Reality Check

Julianna's perspective: 70% of change initiatives fail. The 95% AI failure rate isn't ridiculous when you understand how fundamental shifts in thinking and behavior challenge people.

8:27

AI vs. Automation: The Confusion

Julianna raises the critical question: Are businesses confusing AI with automation? Many who say "we want AI" actually just need automation.

13:13

What Siri Teaches Us About AI

Bert breaks down what happens when you say "Hey Siri, call Bert"—voice recognition, parsing, context lookup, automation. Most of it is automation, not generative AI.

14:47

Addressing the Warrior Archetype

When employees fear AI will take their jobs: "You have 25 years of institutional knowledge. AI might make your reports run faster, but this is where your value lies."

18:01

The "Could vs. Should" Framework

Julianna introduces her friends "could and should." Just because you could implement AI doesn't mean you should. If 65% of your workforce isn't tech-ready, don't start with AI.

19:59

Data Analytics vs. Predictive Analytics

Clear use cases: Data analytics? Use automation. Predictive analytics? That's where AI shines. Understanding which tool to use when.

22:42

Chatbots and Customer Experience

Do you want a chatbot? What will it have access to? A simple FAQ bot vs. one connected to your ordering system and customer data—vastly different use cases.

25:05

Build vs. Buy for AI Tools

Third-party AI tools succeed 67% of the time. Internal builds? Only 33%. The MIT data shows buying specialized vendor solutions works far better than building from scratch.

27:07

Small Business AI Strategy

For small business owners: Could you use AI for sales and marketing content? Yes. Should you? Probably better to hire someone who can actually gain clients.

30:23

The Expense Report Example

AI overkill: Using AI to learn individual spending patterns for expense reports. That's automation territory. Use AI to analyze organizational spending for fraud detection instead.

32:27

When to Use Automation vs. AI

Automation: Repetitive, rule-driven, consistent tasks (compliance, approvals, resets). AI: Ambiguity, pattern recognition, messy data, forecasting, predictions.

36:06

Tech-First Organizations

Reframe the conversation: Instead of pushing "it needs to be AI," ask "Can I use technology to solve this business problem?" Then break down where you need better reporting, automations, or specific types of AI.

Key Quotes

"When you're introducing change, people need a narrative, not a mission statement, not a data dump. A clear human story that says, 'Here's what's changing. Here's why it matters. And here's what it means for you.'"

— Bert Carroll

"I really want to understand the quote unquote generative AI projects that were being done because I have found that a lot of businesses that we've worked with—they're like 'we want AI, we need AI'—and then we sit with them and we're like, you just need automation."

— Julianna Fricchione

"It's kind of like giving your team a map before a hike. You can have the best trail, the most advanced gear, but if no one knows what direction to go, they're just going to stand there."

— Julianna Fricchione

"Let me introduce you to my two friends, could and should. The difference is that we can use AI and we can implement an AI project to solve this problem, but should we do that?"

— Julianna Fricchione

"Let AI do stuff it's good at. Let people do stuff they're good at. Don't make me do with my human meat hands stuff that you're going to be better at."

— Bert Carroll

"If something is repetitive, if something is rule driven, if it's consistent, if it's like a compliance issue, if it's something that you need to audit—that should probably be an automation because it's very black and white."

— Julianna Fricchione

Could vs. Should AI Decision Tree

Use this framework to determine whether you should implement AI or automation for your business challenges

Step 1: Assess Your Organization

Is 65%+ of your workforce tech-ready?

Do they efficiently use computers, Excel, Teams/Zoom, etc.?

❌ If NO → Start with foundational technology training before AI

✅ If YES → Proceed to Step 2

Step 2: Identify Your Problem Type

Use Automation For:

  • • Repetitive tasks
  • • Rule-driven processes
  • • Consistent, predictable workflows
  • • Compliance & audit needs
  • • Approvals, resets, checks
  • • Data analytics (not predictive)

Use AI For:

  • • Ambiguity & nuance
  • • Pattern recognition
  • • Messy, unstructured data
  • • Forecasting & predictions
  • • Predictive analytics
  • • Complex decision support

Step 3: Build vs. Buy

Third-Party AI Tools:

67% success rate according to MIT research

✅ Recommended for most businesses

Internal AI Builds:

33% success rate according to MIT research

⚠️ Only if it's core to your competitive advantage

Key Takeaways

1

Give People a Map, Not a Mission Statement

When introducing change, people need a clear narrative: Here's what's changing, why it matters, and what it means for you. Without this "change script," even the best tools will fail.

2

Could vs. Should

Just because you could implement AI doesn't mean you should. If 65% of your workforce isn't tech-ready, start with foundational technology before jumping to AI.

3

AI vs. Automation: Know the Difference

Automation: Repetitive, rule-driven, consistent tasks (compliance, approvals). AI: Ambiguity, pattern recognition, messy data, forecasting. Use the right tool for the job.

4

Buy Over Build (Usually)

MIT data shows third-party AI tools succeed 67% of the time, while internal builds succeed only 33% of the time. Focus on what you do best, buy specialized tools for the rest.

Ready to Lead Change?

Join the presale for Automate to Accelerate—our self-paced course and live cohort to help you streamline and scale your business with the right technology decisions.

Be the first to know when enrollment opens