Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Data Methodology topic
No spam. Unsubscribe anytime.
Committee hears tri-agency data approach; AI agent groups hundreds of occupations into job functions
Summary
THECB staff described a tri-agency labor-market analysis using SOC/NAICS/O*NET and an AI agent to group ~800 occupations into ~108 job functions before applying labor-market filters to identify viable program-of-study candidates.
Staff presented the data-driven method that will guide program-of-study revisions, including use of SOC and NAICS codes, O*NET job-function extraction and an AI-assisted grouping step.
Valerie described the analytic path: the tri-agency effort aligned roughly 800 occupations into 14 clusters, then extracted job functions from O*NET. An AI agent grouped those functions and produced about 108 candidate job functions; staff reported collapsing seven functions after review and said further filtering with labor-market information (LMI) will reduce the list to programs that meet high-wage, high-skill, high-demand thresholds.
“An AI agent was created that said, okay. Here are the 150 occupations that have been aligned to this career cluster…here are the job functions of all of these occupations. Can you begin to group these?” Valerie said during the presentation. Staff cautioned that this is an analytic tool to inform human-led task groups; the AI output was reviewed and adjusted by THECB and TEA staff.
Committee members asked how LMI will determine which functions fall out and whether regional demand will be respected. Staff said the tri-agency data team is working to ensure the supply side (education and training) is more visible to the analysis and that the AI output is only the starting point for subject-matter deliberations.
The presentation did not include specific numeric LMI thresholds in the meeting record; staff said those filters (high-wage, high-skill, high-demand) will be applied in the next phase of work.
