87% of Finance Professionals Have Ignored Fraudulent Expenses, Reimbursements, or Claims, Medius’ New Financial Census Reveals

A new Medius 2026 Financial Census (surveying 2,386 finance executives in the US, UK, Sweden and France) finds rising tolerance for “shallowfakes” as finance teams invest heavily in AI and automation. Despite this, fraud, operational efficiency gaps, governance issues, and workforce wellbeing remain persistent challenges, suggesting controls and risk management have not kept pace with automation. Overall, the news is a cautious signal for AP automation and spend-management vendors focused on fraud prevention and governance.

Analysis

The market is likely underestimating how quickly AI fraud shifts spend from "automation" to "trust infrastructure." The first-order winners are not the workflow vendors themselves, but the control-plane names that sit between the model and the payment rail: identity, anomaly detection, privileged access, immutable audit trails, and transaction monitoring. That favors cyber/identity exposure more than generic SaaS, because every new automation use case expands the approval surface and raises the cost of a single control failure.

The second-order loser is any finance-automation or AP/spend-management vendor whose sales pitch relies on speed while offering weak proof of control. In the next 1-3 months, the key question is whether buyers treat this as a hygiene issue or as a budget reallocation toward governance-heavy suites; if the latter, smaller point solutions will face longer sales cycles and more procurement scrutiny even if top-line demand stays intact. Over 6-18 months, the real impact is on gross margin and retention quality: vendors that can bundle compliance command better attach rates, while "AI-first" tools without controls may need to discount to win deals.

Contrarian view: the consensus may be too focused on whether AI adoption slows, when the more likely outcome is that adoption continues but becomes more expensive and more concentrated. That means the trade is not a blanket short on automation; it is a relative-value long in security/identity versus a short or underweight in high-multiple workflow software with weaker governance messaging. The thesis is falsified if earnings commentary over the next two quarters shows no elongation in sales cycles and no incremental spend on fraud/identity controls despite continued automation rollout.

Key Decisions for Investors

• Long PANW / CRWD on a 3-6 month horizon, ideally on any 3-5% pullback over the next 1-2 weeks. This is the cleanest way to own the control-layer budget that should benefit if finance teams keep automating while fraud risk rises. Falsifier: no acceleration in identity/fraud-related spend commentary in the next two earnings cycles.
• Pair trade: long OKTA / short BILL into the next earnings season. The relative bet is that identity and access controls get prioritized before additional AP automation spend, especially if buyers demand stronger approval integrity. Exit the short if BILL shows improving net retention or explicit fraud/compliance attach-rate gains.
• Buy CIBR as a basket expression only if it lags on a soft tape; hold 1-3 months. This is a slow-burn controls theme, not a one-day catalyst, so entry matters more than urgency. Reduce if cyber budgets fail to show up in enterprise guidance over the next quarter.
• Do not short generic AI-automation software yet; set an alert instead. Wait for hard evidence of fraud-driven implementation delays, board-level control mandates, or widening losses before betting against the category.
• Watch for any large-cap fintech/software earnings call that mentions longer procurement cycles tied to governance. If that shows up, consider a tactical put-spread on the most expensive workflow name in the tape; absent that, the signal is too anecdotal for a standalone short.