AI Training for HR & People Teams

Where a language model helps an HR function, and where it quietly gets things wrong. Participants sort the drafting and synthesis work from the decisions that need an accountable person behind them.

• How large language models produce a response
• Strengths and failure modes in HR documents and conversations
• Tasks worth assisting, decisions to keep with people
• Hallucination, overconfidence, missing context
• Choosing a low-risk first workflow for an HR team

The group writes prompts it can reuse across People workflows, with context, task, constraints, format, and a check step worked through on the second and third drafts.

• A reusable prompt structure for People workflows
• Supplying useful context without exposing personal data
• Setting audience, tone, reading level, and format
• Asking the model to flag its own assumptions and uncertainties
• Reworking rough output into an approved document
• Starting a governed prompt library for the team

The writing that comes round every quarter gets rebuilt as an AI-assisted flow somebody reviews. Teams design roles and draft handbook or policy content. They also produce learning material a subject-matter owner can check and approve.

• Drafting role profiles from responsibilities and capability needs
• Comparing job descriptions for ambiguity and inconsistent expectations
• Preparing policy and handbook drafts with source references
• Turning expert notes into learning outlines and facilitator guides
• One message, many audiences
• Sign-off: owner review and revision before publication

AI organizes qualitative feedback and supports scenario work without turning model output into a verdict about people. Participants build traceable summaries and question assumptions. They prepare the options for leadership review.

• Grouping survey-comment themes without erasing dissent
• Summarizing interview and listening-session notes with a clear trail back to source
• Observation, interpretation, proposed action
• Mapping skills and role gaps from approved internal inputs
• Building headcount and capability scenarios with explicit assumptions
• Checking whether small groups or minority views vanish from a summary

Recruiting assistance runs against one boundary that never moves: AI can support preparation and administration, but people make hiring decisions. The module covers bias checks and candidate data. It also covers protected handling of employee data.

• Drafting inclusive role adverts and structured interview guides
• Writing candidate communication without inventing commitments
• Who decides: AI supports the process, people make the hiring call
• Testing criteria and outputs for bias and proxy discrimination
• Handling CVs and employee records, including special-category data, safely
• Setting approved tools, access boundaries, retention, and escalation rules

One-off experiments give way to a practice the whole function shares. People leaders leave with pilots, policies, review controls, and learning loops, so responsible use repeats across HR and the wider organization.

• Selecting pilots by value and sensitivity, with reversibility checked
• Naming owners and reviewers, plus escalation paths
• Writing an AI use policy employees can apply
• Quality, adoption, rework
• Supporting managers and teams through role and workflow change
• Building a 90-day rollout plan with review checkpoints