Lessons learned from building training with GenAI
Generative AI has quickly become a powerful partner in learning design. Accelerate early design work by summarizing long interviews with subject matter experts (SMEs), draft learning materials, and content structure. In many ways, they function like energetic research assistants, helping to turn raw expertise into learning experiences.
But anyone who has used generative AI in a real-world project knows the other side of the story. That is, AI is not neutral. If the data is incomplete or the prompt is ambiguous, the system will not simply respond with “I don’t know.” Instead, fill in the gaps. Sometimes they contain plausible but incorrect information. They may fabricate references, draw unsupported conclusions, or confidently propose ideas that do not match the actual context. This is a key challenge for learning and development (L&D) professionals. How can you effectively use generative AI without losing control over the accuracy, reliability, and accountability of your learning content?
In my recent work developing leadership training programs, I’ve found that the answer goes beyond simply improving prompting. The key is to create a process that responsibly integrates AI across learning design workflows. Here are some practices that helped us keep AI productive while staying in control as we built our leadership training.
Grandma’s rule: Always start with the end.
My first rule comes from something I learned long before AI existed. During my high school teacher certification program, my supervisor often repeated this simple advice: “Always start with the goal.” People are different, days are different, and environments are always changing, but goals remain the foundation for keeping the learning experience focused and meaningful. This principle becomes even more important when designing with AI.
Define your learning objectives clearly and clearly before creating your content. All prompts, summaries, and content drafts are tied to these objectives. As your project progresses and your conversations with small businesses deepen, your goals may change slightly, but they will always remain the anchor of your process.
This practice helps prevent a common problem with generative AI: directionless content expansion. AI can generate a large amount of sophisticated material, but without a clear goal structure, that material can move away from learning objectives.
Goals act as a control system that keeps the AI’s output aligned with the training objectives.
Create your own AI assistant using sources
Another important practice is to create project-specific AI assistants rather than relying on generic chatbots. In my workflow, I upload the main materials like this:
Compliance and policy documentation. SME notes and summaries. learning framework. A document that defines the goals and objectives of the course.
These materials become the source base that the AI assistant references when generating content. This approach significantly reduces illusions because the system is guided by verified internal information rather than relying on common Internet patterns. It also keeps the prompts focused and ensures that the material produced is relevant to the specific learning context. In essence, the assistant becomes a structured knowledge environment rather than a free-floating text generator.
Practical practice comes first
One of the most valuable lessons I learned is that authentic learning experiences have to come from real practice, not from the imagination of an AI. Generative AI can create convincing scenarios, but struggles with subtleties such as local language, tone, and professional nuances. These elements are important in leadership training and workplace learning.
To address this, start with real-life experience.
In my projects, educators and facilitators often record short reflection videos for professional development. These videos capture real conversations, authentic language, and the subtle movements of practice. I collect transcripts from these recordings and use them as source material for my AI assistant. You then instruct the AI to generate scripts or scenarios based on those transcripts based on your learning objectives.
This process allows AI to structure and refine materials while preserving the authentic voice of the practitioner. The result is learning content that is natural and grounded, rather than artificial.
Scale learning without losing meaning
One of the most promising uses of AI in learning design is knowledge augmentation. Once your content is based on real-world experiences and aligned with your purpose, AI can help refine and expand your content. For example, I often ask my assistant to improve language clarity and apply SEO-oriented language. This makes it easier to search, discover, and navigate learning materials within digital platforms.
However, this step always occurs after content adjustments, not before. All revisions are double-checked against the learning objectives to ensure that the intended meaning is not distorted by increased clarity or keyword optimization. AI can amplify language patterns, but instructional designers must remain responsible for maintaining the integrity of learning messages.
AI as a language mirror for learning
In Language Machines, Leif Weatherby explains how AI can surface collective linguistic patterns and influence cultural meaning. In many ways, this is exactly what we see when generative AI is used in learning design. AI reflects how people speak, write, and structure ideas across disciplines. When used responsibly, it can help uncover patterns in an organization’s knowledge and accelerate the translation of expertise into learning experiences. However, this only works if AI is carefully integrated within the learning design process.
For me, this means integrating AI across the ADDIE model, analysis, design, development, implementation, and evaluation stages while maintaining strong collaboration with subject matter experts. AI is not going to replace learning designers or small businesses. Instead, it becomes a structured partner that helps you organize your knowledge, refine your language, and expand your learning experience. When used in this way, generative AI does not dilute its credibility. In fact, if we work together wisely, we can help protect it.
