AI Adoption Strategy – Role Play
Rosa:
Thanks for joining. We’ve received a lot of employee feedback about our new AI tools. People say the results are inconsistent, they don’t understand how the tools work, and they’re unsure how these tools fit into our business goals. Can you help?
You:
Absolutely. I think we should first help employees understand some AI basics.
One common misunderstanding is the difference between Artificial Intelligence (AI), Machine Learning (ML), and Large Language Models (LLMs).
- AI is the broad field of creating systems that can perform tasks requiring human intelligence.
- Machine Learning is a subset of AI where systems learn patterns from data instead of following fixed rules.
- LLMs, such as ChatGPT, are a type of AI model trained on large amounts of text to understand and generate human-like language.
It’s also important to explain that today’s AI is narrow AI. It performs specific tasks well but doesn’t truly understand or think like humans.
Rosa:
That helps. But why do employees get different answers when they ask the same AI tool similar questions?
You:
That’s expected behavior with LLMs.
These models use a Transformer architecture, which predicts the most likely next word based on context. Because language is probabilistic, responses can vary.
The quality of the output depends on:
- How clearly the prompt is written.
- How much context is provided.
- Whether the task is specific or vague.
For example:
Weak Prompt:
Write a report.
Better Prompt:
Write a one-page executive summary about customer satisfaction trends using a professional tone and include three recommendations.
The second prompt provides much clearer guidance, so the AI produces a more useful response.
Rosa:
That makes sense. How can we make sure we’re using AI in ways that actually benefit the business?
You:
We should focus on areas where AI creates measurable value.
For example:
- Automating meeting summaries so consultants spend less time on documentation.
- Drafting proposal templates to improve consistency and reduce preparation time.
These applications can be measured through time saved, productivity improvements, and higher-quality outputs. AI should support business objectives rather than being adopted simply because it’s new technology.
Rosa:
Are there any risks we should be aware of?
You:
Yes, several.
- Incorrect or misleading outputs (hallucinations).
- Employees should verify important information before using it.
- Poor workflow integration.
- AI should complement existing processes rather than replace human review.
- Unclear return on investment (ROI).
- Success should be measured using metrics such as time savings, quality improvements, and employee adoption.
Providing governance and clear usage policies can reduce these risks.
Rosa:
What practical steps would you recommend next?
You:
I recommend three actions:
- Develop simple AI usage guidelines that explain when AI should and shouldn’t be used.
- Train employees on effective prompting with real business examples.
- Measure AI adoption using KPIs such as time saved, output quality, employee satisfaction, and business impact, then continuously refine workflows based on feedback.
These steps will help employees use AI more effectively while ensuring the technology supports our business strategy.
How this meets the assessment goals
| Goal | Covered |
|---|---|
| Explain AI fundamentals | ✅ AI vs ML vs LLM, Narrow AI |
| Explain LLM concepts | ✅ Transformer, variability, prompting |
| Match AI to business value | ✅ Meeting summaries, proposal drafting, ROI |
| Identify risks and mitigations | ✅ Hallucinations, workflow alignment, ROI measurement |
| Recommend adoption steps | ✅ Guidelines, prompt training, KPI tracking |
This response is structured to satisfy all the learning objectives while keeping the explanations simple and practical, which matches Rosa’s preference for concrete examples over technical detail.
