Job Description Development & Updates

Close-up of hands typing on laptop sitting at an office desk.

Problem Statement

Creating efficiency and increasing creativity in wording of job descriptions to keep them current, inclusive and aligned with evolving skills.

Description

Using supported AI tools loaded with University of Michigan job description templates, guidelines, and high-quality examples can significantly streamline and strengthen job description development. Rather than starting from scratch, HR teams can generate a strong first draft that follows U‑M standards, reflects the role’s purpose and essential responsibilities, and uses clear, engaging language.

This approach also helps keep job descriptions current and future-ready by suggesting updated skills and competencies as work evolves (e.g., digital fluency, collaboration across distributed teams, data-informed decision-making). With built-in reference materials, AI can propose more inclusive, accessible wording, reduce unnecessary jargon, and promote consistency across similar roles—supporting equitable hiring practices and clearer expectations for candidates and employees.

HR practitioners remain responsible for review and final decisions, using their judgment to confirm accuracy, ensure alignment with unit needs and classification requirements, and tailor language to the role and context. Overall, the use of AI in this workflow improves efficiency, enhances quality, and enables HR teams to focus their time on higher-value work such as consultation with hiring managers, role design, and ensuring alignment with organizational priorities.

Learn More

Contact

Email UHR Generative AI Governance for more information or assistance in any AI in HR needs.

Details

Employee Lifecycle

  • Career Development
  • Recruitment / Hiring

Keywords

  • Hiring
  • Job Descriptions
  • Management

Audience

  • Hiring Managers
  • HR Practitioners

AI Tool Options

  • Gemini
  • NotebookLM
  • UM GPT
  • UM Maizey

Complexity to Develop

Low

Complexity to Use

Low