Introduction
Financial institutions and fintechs spend an estimated $304 billion globally each year on anti-money laundering (AML) and know-your-customer (KYC) compliance, according to research by Ronald F. Pol. Despite this scale of investment, RegTech firm Duna has found that roughly 30% of businesses abandon the onboarding process before completion, raising questions about whether compliance spending is translating into effective, revenue-supporting onboarding rather than simply higher operating costs.
This matters to compliance officers, chief operating officers and executives at banks and fintechs, because business customers typically carry high lifetime value, meaning onboarding functions as a revenue driver as much as a regulatory obligation. When onboarding friction causes legitimate business customers to abandon the process, the cost is not only operational; it represents lost revenue that compliance spending was, in principle, meant to help capture rather than deter.
What Business Onboarding Compliance Involves
Business onboarding compliance refers to the AML and KYC processes financial institutions and fintechs must complete before activating a business customer, including identity verification, document collection, ownership disclosure and risk screening. According to Duna, compliance functions in regulated industries sit at the front line against fraud and regulatory penalties, but manual processes often consume analyst capacity that would otherwise be available for genuine high-risk investigation.
Compliance as a Hygiene Function, Not a Growth Driver
Duna’s analysis frames compliance as what it terms a hygiene factor: once mandatory checks are satisfied, the remainder of the onboarding journey should convert legitimate businesses into active customers without generating unnecessary manual work for either party. In practice, Duna’s research suggests this handoff between compliance and conversion is frequently mismanaged.
How Onboarding Friction Drives Customer Drop-Off
Business customers generally expect onboarding to be fast, yet are often asked for additional documentation, further questions or repeated manual reviews. According to Duna, financial institutions lose 15% of onboarding customers with each additional follow-up request, a loss that compounds as additional requests accumulate. Duna also found that if a business is not onboarded within 24 hours, conversion falls by approximately 55%.
Where the Friction Originates
Much of this friction stems from how compliance work is operationally managed rather than from the substance of regulatory requirements themselves. Duna’s findings indicate that documents frequently sit in inboxes, spreadsheets and shared drives, with requests travelling by email and case reviews passing manually between staff members, a workflow structure that adds time without necessarily improving risk detection.
Key Factors Behind the True Cost of Compliance Operations
Software licensing represents only a fraction of total compliance spending. Hidden beneath that figure are internal operating costs, including staff time spent reviewing and approving customers, along with business process outsourcing (BPO) expenses. Analysts also spend considerable time reviewing cases that ultimately require no action.
False-Positive Rates in Screening
False-positive rates in know-your-business (KYB) screening can exceed 90%, a figure corroborated by research from BCG, McKinsey, PwC and ACAMS. Duna’s own internal calculations place this figure closer to 99%. Financial analysts note that the conventional response to this workload, hiring additional analysts or outsourcing operations, adds cost to an already inefficient process rather than resolving its underlying structure.
How AI Is Changing the Compliance Operating Model
According to Duna, AI-driven compliance systems shift onboarding work out of email and spreadsheet-based workflows and into a structure that collects evidence, applies policy programmatically, and routes cases automatically. In this policy-driven model, compliance policies are converted into code and applied as evidence arrives, rather than interpreted manually by staff from a policy handbook, with only cases requiring genuine human judgement flagged for analyst review.
Reported Efficiency Gains
Duna reports that 90% of onboarding cases can now be reviewed in 41 seconds under this model. Reports indicate that one European e-commerce platform using Duna’s AI-native onboarding system reduced SME onboarding time from eight days to under one minute, cut follow-up case volume by 53%, and lowered drop-off rates by 37%. These figures represent a single client case study rather than sector-wide averages.
Costs, Impact and Adoption Challenges
Duna estimates that only around 10% of firms have made meaningful progress implementing AI within compliance operations, a slower adoption rate than the potential efficiency gains might suggest. This is partly attributed to the requirement that AI-driven compliance decisions remain explainable, auditable and consistent enough to satisfy both internal risk teams and external regulators, a governance bar that is more demanding than typical automation deployments outside regulated industries.
The Ongoing Cost of Reboarding
Onboarding compliance obligations extend across the full customer lifecycle rather than ending at initial approval. Ownership changes, adverse media findings, revised internal policies or new product applications can all trigger reboarding, repeating much of the original manual compliance workload. According to Duna, while most executives track initial onboarding performance, considerably fewer are able to quote their reboarding completion rate or its associated cost.
Risks and Limitations
The efficiency figures cited throughout this analysis, including the 41-second review time and the case study results from the unnamed European e-commerce platform, derive from Duna’s own reporting and a vendor-selected case study rather than independent, third-party verification across a broader sample of institutions. The false-positive rate range, while corroborated by multiple consulting and industry bodies, varies by institution and screening methodology, and Duna’s higher estimate of 99% has not been independently replicated by the other cited sources. Additionally, the requirement for AI-driven compliance decisions to be explainable and auditable to regulators remains a developing area of supervisory practice, and institutions adopting these systems should not assume regulatory acceptance is uniform across jurisdictions.
Future Outlook
Duna suggests that three metrics are becoming central to how chief operating officers should evaluate onboarding performance: the percentage of onboarding volume processed without manual intervention, known as straight-through processing; the true cost per completed business customer, inclusive of operating costs beyond software; and the business onboarding conversion rate. Duna reports that its enterprise customers have seen conversion rates rise by 35% to 38% within six months of implementation, though whether these gains generalise across the wider market remains to be independently confirmed as AI-driven onboarding adoption expands beyond its current, relatively early stage.
Conclusion
Global spending on AML and KYC compliance has reached $304 billion annually, yet a substantial share of business onboarding attempts still fail to convert, pointing to a structural gap between compliance investment and customer retention. Industry data suggests AI-driven, policy-based onboarding models can meaningfully reduce onboarding time and drop-off rates, though current adoption remains limited, and the reported efficiency gains are drawn primarily from vendor-reported case studies rather than broad, independently verified industry benchmarks.
