State of AI in Finance | 2Q26

Executive summary:

• Usage is mainstream, 83.9% of survey participants use AI multiple times daily.
• Time savings are meaningful: 67.9% of respondents report saving at least 3 hours per week.
• At-work penetration is led by Claude (64.3%) and ChatGPT (62.5%), followed by Copilot (45.5%) and Gemini (29.5%).
• Top use cases are Research & Analysis (94.6%), Meeting Prep/Summaries (71.4%), Writing Memos (67.9%), and Data Extraction (60.7%).
• When asked to rank their favorite AI, Claude is the clear #1 product. ChatGPT remains heavily used and usually ranks second, while Copilot shows strong distribution but materially weaker sentiment.
• The most common value proposition for finance professionals is speed: respondents repeatedly frame AI as a force multiplier for research, memo drafting, quick information pulls, and repetitive formatting work.
• The most repeated complaints from finance professionals are hallucinations, weak source fidelity, clunky Excel / PowerPoint output, and the inability to safely use real client or internal data.
• Non-negotiables for finance professionals are centered around accuracy, compliance, privacy, source citations, and tight integration into Excel, PowerPoint, SharePoint, internal drives, and financial data systems.
• In terms of firm policy on AI, many employees are allowed to use only approved enterprise tools, with clear restrictions on confidential data. Many firms are still in pilot mode or channeling usage into internal wrappers.
• Regarding potential headcount reductions, it's less about entire workforce elimination and more about leveraging a smaller headcount to do more: respondents most often expect fewer analysts, associates, operations, and processing-oriented roles per team.