AI FOR HR WORKSHOP: WORKSHOP EXAMPLE MATERIALS
Howden Minneapolis with Josh Lemon, Wednesday, September 16, 2026

These are workshop example materials for the guided build of a Job Description Cleaner
and Job Architecture Matching Tool. They are examples for learning, not any employer's
policy or validated job architecture. They contain no employee personal data.

FILES

1. Job Family Descriptions.csv (50 rows)
   Columns: job_family_code, job_family_name, job_family_description
   Use: load every row into an editable backend table. This is the matching source for
   job family.

2. Job Leveling Guide.csv (15 rows)
   Columns: career_stream_code, career_stream_description, job_level_code, job_level_description
   Streams and levels: SUP S1 to S4; PRO P1 to P6; MGT M1 to M5.
   Use: load every row into an editable backend table. This is the matching source for
   career stream and level.

3. Example Job Description.pdf (3 pages)
   A clean Senior Compensation Analyst job description. Sections: Internal Job Information;
   Job Summary; Key Responsibilities; Required Qualifications; Preferred Qualifications;
   Success Profile; Core Competencies; Physical and Work Environment.
   Use: a reference for structure and writing quality only. It is not messy input and it is
   not the answer key for matching. Do not copy its responsibilities, credentials, or job
   code into other roles.
   Note: the PDF shows job code HRTR-Pro-04 and "Level 4 - Senior Professional". The CSVs
   use HR-TRW, PRO, and P4. Use the PDF for format and writing quality. Use the CSV codes
   and descriptions for matching. Do not auto-copy the PDF level or invent a mapping.

OPTIONAL EXTRA TEST INPUT (separate download on the site)
   messy-job-description.txt: a deliberately untidy fictional posting for testing the
   cleaner once the tool runs. It is test input only, not a source reference.

HOW TO USE TODAY
- Download all three files.
- Open your usual approved chat assistant, attach the three files, and think the tool
  through before writing the Lovable prompt (see Build Together on the workshop site).
- Ask the assistant for a Lovable prompt that uses Lovable Cloud for AI and backend.
  Review it, then paste it into Lovable and attach the same three files there.
- Verify the backend loaded 50 family rows and 15 leveling rows.
- Everything the tool produces is advisory. A person makes the placement decision.