A faster return can hide a later discovery
Salesforce’s survey reports a shorter return under lighter oversight and a longer wait to find agents outside their limits.
A faster reported return can hide a slower discovery. Salesforce reports a short payback beside a longer wait to find agents outside their limits.
Return on investment means reported business gains measured against costs. Governance means the rules, reviews, and records that limit what an agent may do.
Salesforce says its double-blind survey covered 2,025 agent decision-makers across 20 countries. Salesforce fielded it in May 2026.
Among production users, Salesforce reports an average of about eight months to meaningful return on investment. That average starts one clock. The oversight comparisons expose another.
The shorter return came with later discovery
Salesforce reports that groups with lighter oversight reached positive return in 7.2 months, against 9.3 months for heavier governance.
Below-average governance groups found an out-of-bounds agent only after a consequential error in 32 percent of cases. The figure was 18 percent for above-average governance groups.
The newsroom page reports these two governance comparisons:
| Measure | Lower oversight | Higher oversight |
|---|---|---|
| Time to positive return | Lighter oversight: 7.2 months | Heavier governance: 9.3 months |
| Discovery after a consequential error | Below-average governance: 32% | Above-average governance: 18% |
That gap does not price oversight. The late-discovery split blocks that interpretation.
A return date fits on a slide. An undiscovered failure has no date yet.
The survey cannot assign the cause
This study is a vendor survey of self-reported outcomes. It shows association rather than causation. It cannot establish that lighter governance caused the shorter reported return.
Self-reporting can shift both clocks. A team decides when gains become meaningful, and it can report a failure only after someone finds it.
The survey compares group-level results. It does not follow one controlled intervention that removes oversight while holding data, scope, workflow, and measurement constant.
Salesforce and its respondents are neither Muniment customers nor endorsers.
Muniment’s conclusion is narrower than the survey headline. Time to a reported return and time to a discovered failure are different clocks.
Operators should ask what remained undiscovered on the date a return appeared. Faster reporting loses its shine when the second clock has barely started.