The AI Adoption Plateau: Why Change Management Still Rules Everything
FutureCraft GTM
Navigating AI Challenges in Content Creation
This chapter addresses the complexities of using AI for content creation, especially the issue of hallucinations and strategies for quality control. It also explores interactive methods to transform traditional content formats into engaging experiences, demonstrating the potential of AI in enhancing team collaboration.
In this episode of the FutureCraft GTM Podcast, hosts Ken Roden and Erin Mills reunite with returning favorite Liza Adams to discuss the current state of AI adoption in marketing teams. Liza shares insights on why organizations are still struggling with the same human change management challenges from a year ago, despite significant advances in AI technology. The conversation covers practical frameworks for AI implementation, the power of digital twins, and Liza's approach to building hybrid human-AI marketing teams. The episode features Liza's live demonstration in our new Gladiator segment, where she transforms a dense marketing report into an interactive Jeopardy game using Claude Artifacts.
Unpacking AI's Human ChallengeLiza returns with a reality check: while AI tools have dramatically improved, the fundamental challenge remains human adoption and change management. She reveals how one marketing team successfully built a 45-person organization with 25 humans and 20 AI teammates, starting with simple custom GPTs and evolving into sophisticated cross-functional workflows.
- The Digital Twin Strategy: Liza demonstrates how creating AI versions of yourself and key executives can improve preparation, challenge thinking, and overcome unconscious bias while providing a safe learning environment for teams.
- The 80% Rule for Practical Implementation: Why "good enough" AI outputs that achieve 80-85% accuracy can transform productivity when combined with human oversight, as demonstrated by real-world examples like translation and localization workflows.
- Prompt Strategy Over Prompt Engineering: Liza explains why following prompt frameworks isn't enough—you need strategic thinking about what questions to ask and how to challenge AI outputs for better results.
Edited Transcript:
Introduction: The Balance Between AI and Human Skills
As AI democratizes IQ, EQ becomes increasingly important. Critical thinking and empathy are important, but I believe as marketers, balance is actually more important.
Host Updates: Leveraging AI Workflows
Ken Roden shares his approach to building better AI prompts by having full conversations with ChatGPT, exporting them to Word documents, then using that content to create more comprehensive prompts. This method resulted in more thorough market analysis with fewer edits required.
Erin Mills discusses implementing agentic workflows using n8n to connect different APIs and build systems where AI tools communicate with each other. The key insight: break workflows down into steps rather than having one agent handle multiple complex tasks.
Guest Introduction: Liza Adams on AI Adoption Challenges
Liza Adams, the AI MarketBlazer, returns to discuss the current state of AI adoption in marketing teams. Despite significant technological advances, organizations still struggle with the same human change management challenges from a year ago.
The Core Problem: Change Management Over Technology
The main issue isn't about AI tools or innovation - teams can't simply be given ChatGPT, Claude, Gemini, and Perplexity and be expected to maximize their potential. Marketing teams are being handed tools while leaders expect employees to figure out implementation themselves.
People need to see themselves in AI use cases that apply to their specific jobs. Joint learning sessions where teams share what works and what doesn't are essential. The focus has over-pivoted to "what's the right tool" when it should be on helping people understand, leverage, and make real impact with AI.
The AI Adoption Plateau
Many organizations face an AI adoption plateau where early adopters have already implemented AI, but a large group struggles with implementation. Companies attempting to "go fully agentic" or completely redo workflows in AI are taking on too much at once.
Success Story: The 45-Person Hybrid Team
Liza shares a case study of a marketing team with 45 members: 25 humans and 20 AI teammates that humans built, trained, and now manage. They started with simple custom GPTs, beginning with digital twins.
Digital Twin Strategy for AI Implementation
Digital twins are custom GPTs trained on frameworks, thinking patterns, publicly available content, and personality assessments like Myers-Briggs. These aren't designed to mimic humans but to learn about them and find blind spots, challenge thinking patterns, and overcome unconscious bias.
For executive preparation, team members use digital twins of leadership to anticipate questions, identify gaps in presentations, and prepare responses before important meetings.
The progression: Simple digital twins → Function-specific GPTs (pitch deck builders, content ideators, campaign analyzers) → Chained workflows across multiple departments (marketing, sales, customer success).
Prompt Strategy vs. Prompt Engineering
Following prompt frameworks (GRACE: Goals, Role, Action, Context, Examples) isn't enough if the underlying thinking is basic. AI magnifies existing thinking quality - good or bad.
Example: Instead of asking "How do I reduce churn?" ask "Can you challenge my assumption that this is a churn problem? Could this data indicate an upsell opportunity instead?"
This transforms churn problems into potential revenue opportunities through different strategic thinking.
The 80% Rule for Practical AI Implementation
AI outputs achieving 80-85% accuracy can transform productivity when combined with human oversight. Example: A team reduced translation and localization costs from tens of thousands of dollars monthly to $20/month using custom GPTs for eight languages, with human review for the final 15-20%.
Measuring AI ROI: Three Strategic Approaches
- Align with Strategic Initiatives: Connect AI projects to existing company strategic initiatives that already have budgets, resources, and executive attention.
- Focus on Biggest Pain Points: Target areas where teams will invest resources to solve problems - excessive agency costs, overworked teams, or poor quality processes.
- Leverage Trailblazers: Identify curious team members already building AI solutions and scale their successful implementations.
Handling AI Hallucinations and Quality Control
AI models hallucinate 30-80% of the time when used as question-and-answer machines for factual queries. Hallucinations are less common with strategic questions, scenario analysis, and brainstorming.
Prevention strategies:
- Limit conversation length and dataset size to avoid context window limitations
- Use multiple AI models to cross-check outputs
- Implement confidence checking: Ask AI to rate confidence levels (low/medium/high), explain assumptions, and identify what additional information would increase confidence
Live Demo: Claude Artifacts for Interactive Content
Liza demonstrates transforming the 2025 State of Marketing AI report into an interactive Jeopardy game using Claude Artifacts. The process involves uploading a PDF, providing specific prompts for game creation, and generating functional code without technical skills.
This "vibe coding" approach allows users to describe desired outcomes and have AI build interactive tools, calculators, dashboards, and training materials.
Future of Marketing Jobs and Skills
Emerging roles: AI guides, workflow orchestrators, human-AI team managers Disappearing roles: Language editors, basic researchers, repetitive design tasks
Transforming roles: Most existing positions adapting to include AI collaboration
Critical skill for the future: Balance
- Innovation with ethics
- Automation with human touch
- Personalization with transparency
Balance may be more important than emotional intelligence as AI democratizes cognitive capabilities.
Key Takeaways
The Gladiator segment demonstrates how dense research reports can become engaging, interactive content without engineering resources. Making AI implementation fun helps teams stay balanced and avoid overwhelm.
Success comes from starting with tiny AI wins rather than comprehensive strategies, focusing on human change management over tool selection, and building systems that augment rather than replace human creativity.
This version removes the conversational back-and-forth while preserving all the searchable content people would look for when researching AI implementation, digital twins, prompt strategy, change management, and practical AI use cases.
Stay tuned for more insightful episodes from the FutureCraft podcast, where we continue to explore the evolving intersection of AI and GTM. Take advantage of the full episode for in-depth discussions and much more.
----more----
To listen to the full episode and stay updated on future episodes, visit the FutureCraft GTM website.
Disclaimer: This podcast is for informational and entertainment purposes only and should not be considered advice. The views and opinions expressed in this podcast are our own and do not represent those of any company or business we currently work for/with or have worked for/with in the past.
The AI-powered Podcast Player
