“I asked the AI to draft our department’s update, but the output was completely useless and full of generic jargon.”
“My team member just sent me a massive, unedited block of text with the note, ‘Gemini told me this,’ leaving me to sort through the mess.”
“I am spending more time fixing poorly prompted AI outputs than it would have taken to just write the document myself.”
These frustrated refrains are becoming more common in everyday workplace conversations, pointing to a much larger and more complex discussion taking place across the University. Even with the rise of advanced “agentic AI,” the reality is that most of us are primarily using AI to look up information, learn how to do something, or to support everyday communication such as drafting emails, revising documents, or summarizing information. And that is where many frustrations begin.
There is a growing divide in the workplace. On one side is “low-value AI use”: the “one-step copy-paste.” This happens when someone enters a quick prompt, takes the immediate result, and passes it along without review. This kind of direct, uncritical use is what often fuels concerns about overreliance, lower-quality thinking, and a loss of professional judgment.
On the other side is constructive, higher-value use. In this approach, employees use structured prompts to build scaffolding for their work. They bring in their own expertise, use multi-step workflows, check facts, revise for the audience, and consciously remove the generic “AI tone” so the final product sounds human, useful, and appropriate.
Moving Forward: Evaluating Your Team’s AI Literacy
To move forward, it can be helpful to understand where your team, department, or individual employees currently sit on the spectrum of AI literacy. People are at very different stages of familiarity, comfort, and skill with these tools.
Adapted from: Rintz, W. (2025). Develop Organizational AI Literacy in the Age of Generative AI. Presented at the 2026 Association for Talent Development International Conference, Los Angeles, CA.
- Level 0: No AI Adoption. Employees avoid using the technology entirely and maintain traditional, manual ways of working.
- Level 1: Copy-Paste / Direct Use. This represents many users today. They enter a single, basic prompt without a systematic approach, most often using the tool to refine communication drafts such as emails or standard documents.
- Level 2: Constructive / Indirect Use. Employees use structured prompts to build a foundation for their work. They use the tool to improve their thinking, organize ideas, test options, and strengthen problem-solving.
- Level 3: Workflows. Users intentionally embed AI as a specific, planned step within an existing institutional workflow or departmental process.
- Level 4: Agentic Use. Advanced users create customized agents or sets of agents to support more complex workflows.
Take a look at where you and your team sit on this spectrum, and consider what the next step makes sense for your specific work.
If your team is not ready to use generative AI, that does not mean you are behind. You can still make meaningful progress by improving your existing workflows, clarifying decision points, and cleaning, structuring, or organizing the data your department uses. These steps help build a stronger foundation for future AI use.
If your team members are newer to generative AI, encourage them to experiment with their prompts in small, focused ways. Try changing just one or two variables, such as the role you assign the tool, the audience, the tone, or the boundaries. For example, a response can shift dramatically when the system is asked to act like an entry-level assistant versus an experienced specialist. These small experiments are a form of instruction, context-setting, and revision.
For employees who are already comfortable with prompting, the next step is often scaffolding the work more intentionally. This may mean bringing in deeper, more localized context through tools such as Gemini Gems or NotebookLM, where the AI can work from relevant documents, examples, or organizational information. This stage may not always feel faster at first. In fact, it can feel less efficient while people are learning how to structure the process. But users at this level are often able to produce more thoughtful, higher-quality outputs because they are using the tool to expand their thinking, not replace it.
Ultimately, the goal is not to outsource thinking to a large language model. The goal is to identify which parts of your department’s work might be strengthened by it. Look closely at your workflows and determine where AI can reasonably help: drafting, summarizing, organizing ideas, identifying patterns, preparing scenarios, or testing language. Then ask the most important question: Does this use of AI improve the quality of the work?
AI may change the tools we use, but it does not change the need for judgment, accountability, and thoughtful supervision. In many ways, it makes those responsibilities even more important.
