Why the onboarding-speed race is the wrong race – and what knowing your customer is actually for.
For a decade, client lifecycle management and KYC have been pictured as a speed problem: shorten cycle times, strip out manual work, stop losing clients to a months-long account-opening slog. That story was never wrong, but it was always small. It optimized the cheapest, least consequential moment in the entire client relationship. And it did so for an entirely legitimate reason: growth. Onboarding is the moment the relationship – and the revenue – actually begins, so every day shaved off it is a day of earlier returns and one fewer reason for a client to walk to a competitor with a smoother front door. In a market where client experience became a genuine differentiator and time-to-revenue a board-level metric, betting the KYC budget on onboarding was a rational commercial decision.
But the growth case quietly became the only case.
And so, the market has spent its AI budget the same way – layering models onto the same onboarding-then-periodic-review workflow so analysts can clear cases faster. It is worth seeing that for what it is. Accelerating the old model leaves the old model intact: the headcount, the silos, and the architecture all survive; they simply run quicker.
Which is why the more interesting question is not how do we do KYC faster? It is closer to: What if we reconsidered what knowing a customer even means? Start there – and start by looking at where the enforcement actions actually land.
Regulators rarely fine an institution for onboarding a client too slowly. They fine it for failing to notice what that client became – the ownership structure that quietly changed hands, the dormant entity that lit up with unexpected flows, the periodic review that rubber-stamped a profile no one had truly looked at in three years. The risk, the cost, and the liability all live downstream of the front door. Onboarding is a one-time event. Monitoring is forever.
The more honest reframing is not, “How do we onboard faster?” It is “How do we make, govern, and evidence client-risk decisions consistently across an entire lifecycle that never stops moving, without compromising on speed and experience?” Once you ask the question that way, client lifecycle management and KYC stop being a journey to optimize and become something architecturally more serious – the control plane through which every risk decision is made, revisited, and justified over time.
Knowing the customer is Business 101
Strip away the regulatory framing for a moment. Knowing who you are dealing with – before you engage and for as long as you engage – is not a compliance task bolted onto the side of the business. In a real sense it is the business: It is how a firm decides whether to take someone on, how to price them, how to serve them, and how to protect its own name. Compliance-grade KYC is one expression of a larger question, “Should we do business, and on what terms?”, not a separate chore running in parallel to it.
That reframing matters because it changes what you are optimizing. If KYC is a cost center, the goal is to make the cost smaller and the queue shorter. If we look at KYC as the mechanism by which the business decides who it deals with, the goal is to make that judgement sharper, faster, and continuously current. The first framing yields a quicker version of the old process, while the second redefines the process around the thing the business actually needs to know.
Next in the series →
If the goal is not speed, what is it? In Part 2, “Banks Bought Data. Regulators Examine Decisions.”, we look at the one thing supervisors actually examine. Spoiler alert: It is not your data lake, and it is not your model’s accuracy. It is the decision itself, and why most architectures can produce an answer they cannot defend.