Example verification audit for demonstration purposes. All companies and data are fictional. Methodology is real.

73%

Verified

Verification Audit: AI-Generated PE Investment Memo

A private equity firm submitted a 2-page AI-generated due diligence summary on a South African fintech target for verification before presenting to the investment committee.

73%

Claims verified

82%

Citations valid

58%

Assumptions sourced

Document Summary

Atomic claims extracted

Verified against primary sources

Flagged as unsupported or fabricated

Partially supported (overstated)

Source types used for verification:

8 financial filings5 industry reports3 regulatory documents1 academic paper

Claim-by-Claim Analysis

Showing 6 of 23 extracted claims. Each claim is verified independently.

“The South African fintech market grew at a CAGR of 18.2% between 2020 and 2024”

Verified

Disrupt Africa's Fintech Landscape 2024 report confirms an 18.2% CAGR for the SA fintech sector across the measured period, consistent with GSMA mobile money growth data.

Source: Disrupt Africa, Fintech Landscape 2024, p.34

Prior: 65%→Posterior: 94%

“The target company processes over R2.4 billion in monthly transaction volume”

Verified

Company audited financial statements for FY2024 report average monthly TPV of R2.47 billion. Figure is consistent with quarterly investor updates.

Source: Company financial statements FY2024, Note 14

Prior: 60%→Posterior: 92%

“FSCA has signalled intent to introduce a dedicated fintech licensing framework by Q3 2025”

Fabricated

No FSCA publication supports this claim. The FSCA Regulatory Strategy 2024–2027 mentions fintech oversight under existing frameworks but specifies no timeline and no dedicated licensing framework. The claim appears to be a hallucination.

Source: FSCA Regulatory Strategy 2024–2027 (searched); FSCA media releases 2023–2024 (no match)

Prior: 40%→Posterior: 3%

“Customer acquisition cost declined 34% year-on-year”

Unsupported

No source provided in the memo. Company management presentation (Q4 2024) mentions “significant CAC improvement” but does not quantify. The 34% figure cannot be verified against any available source.

Source: Company Q4 2024 management presentation (partial); no primary source identified

Prior: 45%→Posterior: 28%

“The target holds a 12% market share in South African digital payments”

Overstated

Mastercard/FNB Digital Payments Index 2024 estimates 6–8% market share for the relevant segment (merchant-acquired digital payments). The 12% figure appears to conflate total digital payments with the narrower merchant-acquired segment.

Source: Mastercard/FNB Digital Payments Index 2024, Table 7

Prior: 50%→Posterior: 15%

“Comparable transaction multiples for SA fintech range from 8–14x revenue”

Overstated

Capital IQ data shows 6–12x revenue for the most recent comparable transactions (2022–2024). The upper bound of 14x reflects a single outlier (a cross-border payments acquisition in 2022). Median is 9.2x.

Source: Capital IQ, SA Fintech M&A Comparables 2022–2024

Prior: 55%→Posterior: 22%

Assumptions Audit

Assumption	Value	Source	Status
WACC	15.2%	PwC Valuation Methodology Survey 2024	Verified
Terminal growth rate	5.5%	Sourced to "consensus" — no specific publication cited	Unsupported
Revenue CAGR projection	22%	Company management forecast (FY2025–FY2028)	Source identified (forward-looking)

Findings Summary

4 of 23 claims could not be verified against any primary source.
1 claim cited a regulatory action that does not exist in the public record.
2 claims overstate the evidence by conflating market definitions or relying on outlier data points.

Verdict: The investment memo is usable after corrections. The fabricated FSCA claim must be removed. The market share and CAC claims require sourcing or qualification before presentation to the investment committee.

Methodology

This audit was conducted using brfcase's atomic Bayesian verification engine. Each claim in the submitted document was decomposed into independently verifiable atomic propositions. Claims were searched against FSCA regulatory publications, company financial statements, industry reports (Disrupt Africa, GSMA), Capital IQ transaction data, and academic research via OpenAlex. Prior probabilities were assigned based on claim type and source availability. Posterior probabilities were calculated using evidence weight from matched sources. Full audit trail available on request.

This is an example of a standard verification audit.

To run one on your own document, get in touch.

Request a verification audit