South AfricaFinance and insurance2019 – ongoing

FNB

AI financial-crime investigations and report drafting

FNB said AI freed at least 70% of analysts' time in financial-crime investigations

WorkingCompany's own claim

Problem

Financial-crime analysts collected information across systems and wrote investigation reports, leaving less time for decisions.

Approach

FNB combined models that detected financial-crime patterns with models that generated reports for the Financial Intelligence Centre. It began large-scale use in 2019 with internal small language models and later added trained large language models.

Results

at least 70%of analysts' time freed on average, according to FNBCompany's own claim[1]

more than 160,000investigations processed using AI in the last financial year reported by FNBCompany's own claim[1]

Takeaway

Analysts retained decision-making while AI handled information gathering and report preparation. The time savings were reported by FNB's chief data and analytics officer.

Sources

Banks ramp up spending on artificial intelligence to boost efficiency
Daily Maverick7 Jul 2024News article

Source summaries

[1]Daily Maverick7 Jul 2024

Daily Maverick reported FNB's account of AI-assisted financial-crime investigations. Its executive attributed more than 160,000 investigations and at least 70% of analysts' time freed to the system.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026