Leveraging AI for Research Administration Training

By: Katie Gomez Freeman 

The world of research administration is complex and constantly changing. Whether it’s changing regulations, varied sponsor requirements, or complex financial compliance, every day looks different. This introduces a unique challenge to both onboarding new employees to the field, and ensuring that current employees maintain up-to-date knowledge on core concepts while introducing new guidelines.  

Artificial Intelligence (AI) is one way that we can address the challenges of onboarding, training, and education of research administrators. AI tools are not just for automating tasks but can also be key in revolutionizing how we capture, structure, and disseminate institutional knowledge.  

Ways AI Can Help

  • Creating outlines of training content – Sometimes the hardest part about generating content is simply knowing where to start! AI can be very helpful in providing a starting point as content is developed to ensure there are no gaps or key concepts missing.

Example Prompt

Create a clear and structured outline for a research compliance training program designed for [insert primary audience].

The outline should organize the training into logical modules or sections that progress from foundational concepts to more practical applications. For each module, include:

  • A brief description of the topic
  • Key subtopics to be covered
  • Suggested learning objectives

Emphasize areas of research compliance most relevant to the audience, such as regulatory requirements, institutional policies, sponsor expectations, and common compliance risks or challenges they may encounter in their roles.

The outline should be practical and easy to follow.

  • Generating quiz questions or case studies – AI does a great job of reviewing content you provide and helping generate quiz questions based on the content. It can also create quick and engaging case studies that can be incorporated into training materials. 

Example Prompt

Create an engaging, realistic case study for a training session on [insert compliance topic] designed for research administrators.

The case study should present a brief narrative scenario (1-2 paragraphs) involving a university setting where a [insert compliance issue] emerges during the lifecycle of a sponsored project (e.g., proposal development, award management, or project implementation).

Include enough contextual details (roles of the PI, department staff, and the Office of Sponsored Programs) to make the situation relatable and realistic.

After the scenario, provide 3–5 discussion questions that prompt learners to identify the compliance issue, consider potential risks or consequences, and discuss appropriate actions or institutional responses.

The tone should be practical and thought-provoking, encouraging participants to analyze the situation and apply compliance concepts in a real-world research administration context.

  • Assessing gaps or identifying weak points in existing materials – AI can also be utilized to improve upon existing materials. This may be helpful to identify if there are any knowledge gaps that need to be incorporated or if a certain concept could be clarified.  

Example Prompt

Review the following training or guidance material intended for [insert audience, e.g., new research administrators] on [insert topic] and assess whether there are gaps, unclear explanations, or areas where additional information would improve understanding.

Identify any concepts that may be missing, underdeveloped, outdated, or potentially confusing for the intended audience.

In your assessment, note specific sections where clarification, examples, or additional context could strengthen comprehension, and recommend concrete improvements such as adding explanations, restructuring content, incorporating examples or scenarios, or expanding coverage of key topics.

The goal is to improve the overall effectiveness, clarity, and completeness of the material while ensuring it aligns with the knowledge level and needs of the intended audience. 

  • Creating step-by-step procedures or process recording – Another key aspect of onboarding is being able to review and learn from existing processes. Certain AI tools can be integrated into Zoom or Teams meetings to generate an output with a step-by-step procedure of a certain process, making it a quick way to document procedures. 

Example prompts were generated by ChatGPT and refined by the author

As with any AI use case, it is also important to recognize certain risks that should be kept in mind and addressed in the context of training.  

  1. Outdated regulations – Because of the constantly changing regulations and requirements, AI may have a hard time knowing what the most up-to-date information is. That is why I typically recommend utilizing it as a starting point and/or making sure research administration experts review any outputs before implementing/finalizing the materials.  
  1. Data Privacy – It is also important to remember to never input proprietary institutional data or sensitive PI information into public AI models. 

The key to addressing the risks of utilizing AI for training materials is maintaining a human-in-the-loop approach. Subject matter experts should always validate outputs before finalizing materials to make sure content is accurate and up-to-date.