Expert AI Data Demand Index

Current findings

None. Sample size is zero published records. When a series has enough real observations to report without identifying a client, it will appear here with a date and a limitation note.

Sample size

Not yet published

Time period

Not yet published

Data source

Planned: de-identified inquiry and delivery metadata from Stratum programs, never raw client data.

Last updated

Methodology only · 2026-08-26

Methodology

Future reports will count structured fields from qualified buyer inquiries and, where a program ran, from delivery records: domain, credential class, requested headcount, duration, work type, and whether adjudication was in scope. Counts will exclude spam, incomplete forms, and any record that cannot be de-identified.

The index will describe demand as observed by one company. It will not be a census of the market. Dataset schema will be added only when a public, structured dataset actually exists.

Limitations

• Selection bias: only teams that contact Stratum.
• Small-n risk: early series may be too thin to publish.
• Confidentiality: some high-signal slices may never be publishable.
• No causal claim that demand equals the “right” staffing model.

Planned series

• Most requested domains
• Most requested credentials
• Average requested cohort size
• Project duration
• Evaluation types
• Adjudication frequency
• Geography
• Specialty demand