Acuiti says 44% of respondents in its Q3 2026 derivatives-management research report significant, quantifiable time or cost savings from artificial intelligence. The result, published on 22 September, comes from its Derivatives Expert Network of senior executives across sell-side execution, clearing, proprietary trading, hedge funds and asset management. The quoted 44% measures responses within that research exercise; it is not evidence that precisely 44% of all derivatives firms worldwide have already achieved such savings.[1][2]
The published survey summary reports materially different controls by firm type. All responding banks said they controlled employees' AI use, and two-thirds had a dedicated AI risk committee. Around half of responding proprietary-trading firms allowed relatively free employee use, and about one-third had no formal governance policy. Those contrasts help explain why reported productivity gains do not amount to a uniform industry operating model.[1][2]
The survey release does not, on its own, prove the savings were caused by AI rather than other workflow changes. It also does not disclose a transaction-level benchmark, standardised calculation period for claimed benefits or an independently verified financial return across participants. Self-reported measured savings are stronger evidence than purely aspirational adoption language but weaker than a controlled firm-by-firm before-and-after study.[1][2]
What a brokerage could verify internally
A further 42% of network respondents reported benefits they had not yet fully quantified. That separates measured benefits, as respondents describe them, from perceived improvements still awaiting calculation. The public release does not provide the exact respondent count or enough methodological detail to treat either percentage as a representative estimate for all derivatives firms.[1][2]
Governance should specify who approves model changes, what data can leave the firm, whether client information is retained by an external provider, how outcomes are reviewed, and whether staff can reliably roll back faulty automation. A trading desk's operational risk may rise even when average processing time falls if an AI system can incorrectly classify positions without a human checkpoint. These are implementation considerations, not findings that the Acuiti survey measured each control in every organisation.[1][2]
Interpreting the Acuiti figure
| Measure | Evidence boundary |
|---|---|
| Survey release | 22 September 2026 |
| Respondents | Senior executives in Acuiti Derivatives Expert Network |
| Reported measurable savings | 44% of survey respondents |
| Benefits not fully quantified | A further 42% of respondents |
| Nature of measure | Self-reported time or cost savings |
| Whole-industry adoption estimate | Cannot be inferred from the release alone |
| Independent causal return on AI | Not established |
Acuiti consolidated previously separate buy-side, proprietary-trading, sell-side and clearing research into a broader quarterly report. That design expands topical coverage but makes survey sample and methodology essential before inferring changes relative to earlier reports. The supported headline is that a substantial share of surveyed executives say they have quantified benefits while governance remains heterogeneous; an unsourced estimate of aggregate industry savings would go beyond the available evidence.[1][2]
Sources & further reading
- AI moves from experimentation to governed deployment as productivity gains grow Acuiti · 22 Sep 2026 · Accessed 30 Oct 2026
- Derivatives Management Insight Report Q3 2026 Acuiti · 22 Sep 2026 · Accessed 30 Oct 2026