More than six in ten financial services professionals believe an AI-generated error has reached a client or internal decision-maker in the past year.
The revelation comes from Macabacus, a Microsoft 365 productivity platform for finance and professional services teams based in New York, which released its 2026 GenAI in Financial Services: Velocity and Verification report on September 10, 2026.
The company surveyed its 75,000 users and analyzed conversations with hundreds of clients and prospects across investment banking, private equity, corporate finance, and advisory firms.
Of the 62 percent who believe an error has escaped review, 46 percent described the situation as a "probably, but no one noticed" scenario - an acknowledgment that mistakes may be circulating in client presentations and financial models without anyone catching them. Only 38 percent said they were confident no such error had occurred.
According to the Macabacus report, 87 percent of respondents use AI daily or weekly to generate financial models and client presentations.
However, only 23 percent of firms have comprehensive guardrails in place, defined as approved tools, accuracy checks, brand compliance, and structured review workflows. A further 36 percent of daily or weekly AI users work at firms with no guardrails at all.
The disparity points to a structural problem that wealth management and advisory professionals are beginning to reckon with. Financial advisors and wealth managers are increasingly relying on AI tools to handle everything from portfolio analysis to client communications, but the infrastructure to validate that output has not kept pace.
Paul Ross, chief marketing officer at Macabacus in New York, framed it this way: "Deal teams should not slow down their use of AI. They need guardrails that let them move faster while maintaining accuracy and their clients' trust."
The report also surfaces a meaningful divide between how junior and senior professionals assess the reliability of AI output.
Among analysts and associates, 43 percent said AI has made them more confident in their models and presentations. Among vice presidents, directors, and managing directors, that figure drops to 29 percent - and senior reviewers are six percentage points more likely than their junior counterparts to say AI has made them less confident in the work.
That divergence matters in practice. Junior staff are often the ones generating AI-assisted deliverables, while senior reviewers are tasked with catching problems. If the people closest to the output feel more confident than the oversight layer, errors have more room to pass through undetected.
Firms and advisors navigating the integration of AI into client-facing workflows are confronting this dynamic directly. A portfolio management firm quoted in the Macabacus report put it plainly: "It's a glorified intern. Whatever you would give an intern to do, you're still going to review it after. You're going to proof it and challenge it. You need to do that with AI."
When asked what would give them greater confidence in AI-generated content, respondents were broadly aligned.
Twenty-seven percent cited a full audit trail showing what AI had changed; another 27 percent called for mandatory human review at the handoff stage. Twenty-five percent said they wanted automated verification built directly into the tools they already use.
Only 5 percent said firm-approved tools alone would be sufficient. Taken together, 80 percent of respondents said they wanted some combination of a technology layer, human review, or in-tool checks.
The time cost of the current approach is also significant. According to the report, 85 percent of respondents spend 30 minutes or more checking AI-generated models and content before it goes to a client. That overhead reflects the absence of automated verification - time that better guardrails could redirect toward higher-value work.
The demand signal from the industry is clear. Advisors and finance professionals using AI to accelerate their workflows are not asking for slower adoption.
They want the confidence to move faster, backed by systems that catch what humans miss. As AI tools become more deeply embedded in advisory practice management, firms that build verification infrastructure now are likely to find themselves better positioned to protect both client trust and their own credibility.
The full Macabacus report is available at macabacus.com.
Six-figure losses, social media lures and impostor schemes are draining older clients’ savings, report to Congress shows
William Tunink, an industry veteran, was indicted for running a Ponzi scheme this month in Des Moines.
Broker-dealers will soon share targeted returns under SEC-approved Rule 2210 changes, but one securities lawyer warns the regimes aren’t identical.
Insurers ranking independent broker-dealers ahead of RIAs for near-term growth as fee-only fiduciaries still find annuities a hard fit
Most voters 50-plus in swing districts favor candidates backing a caregiver tax credit, as caregiving drains retirement savings
Knighthead Life entered the market with a competitive MYGA. A strong launch earned advisor confidence and paved the way for FIAs.
As AI makes financial information more accessible than ever, Lana Hock explains why human judgment, trust, and empathy remain the qualities clients value most in a financial advisor
