Designing Digital Onboarding for Confident Customer Journeys
How banks can reduce unnecessary friction, preserve verification and give every onboarding exception an owner, a clear next step and a safe route to resolution.
Perspectives · The Banking Balancing Act
A better beginning includes the next step
Digital onboarding has made opening a banking relationship more convenient. Customers can complete straightforward tasks with fewer visits, and banks can apply consistent checks across a much larger flow of applications. The next opportunity is to make that convenience dependable when a customer’s situation needs more attention.
A name differs across records. A verification service is unavailable. A customer needs assistance. The application creates an account reference, but a required review remains pending. These conditions reveal whether the journey has been designed beyond its quickest route.
The constructive goal is to reduce unnecessary friction while making necessary verification and exceptions easier to understand, own and resolve.
Separate three decisions that often share one screen
Identity, product eligibility and operational readiness answer different questions. Has the bank established the customer’s identity through the applicable process? Is the proposed product permissible for this relationship? Have required controls and setup tasks completed so that the account can be used?
A green result in one step does not establish all three. A customer can be identified but need a different product. An account record can exist while activation remains pending. A technical service failure can interrupt verification without establishing anything adverse about the applicant.
Make these distinctions explicit in the decision model and customer status. Product design should use the applicable rules for the institution, customer and route. An assisted path helps complete those requirements; it does not create an exception to them.
Design the review path as carefully as the straightforward path
Our proposed operating model has three outcomes: proceed when sufficient evidence and required checks support it; stop when applicable rules or established risk require it; and review when an authorised process can resolve uncertainty. This is an editorial design framework, not a regulatory classification.
The review path needs a case reference, a named owner, a specific evidence request, appropriate decision authority and a clear update route. Each handoff should transfer both context and responsibility. A branch referral is much more useful when the branch can retrieve the existing case than when the customer must retell the application history.
Consider a fictional applicant whose name varies across documents. If approved policy permits additional authoritative evidence, the next task is to obtain and assess it. Automatic approval would assume the discrepancy is harmless. Immediate rejection would assume it cannot be resolved. A governed review preserves the bank’s ability to reach a justified decision.
Automation changes the work that remains
Automating routine cases can reduce their handling effort while increasing the average complexity of cases reaching people. The manual queue becomes concentrated in evidence inconsistencies, unusual product conditions and dependency failures. The institution needs to equip that queue for its new mix.
An illustrative capacity example makes the consequence visible. Suppose 10,000 applications arrive in a fictional week. Automation resolves 90%, leaving 1,000 reviews. If each review requires 18 minutes, the remaining work is 300 hours before rework. At an invented capacity of 240 review hours, unresolved work grows by 60 hours. The completion percentage alone does not reveal that gap.
These are hypothetical planning inputs, not industry benchmarks. They show why a programme should measure exception volume, complexity, quality and sustainable capacity together. A higher straight-through rate can be valuable while a smaller unresolved queue still needs more specialist skill.
Resume capability needs evidence discipline
Preserving progress saves effort. It also creates an important boundary: the fact that a screen was completed does not mean its evidence remains usable indefinitely. Record the evidence source, verification time and applicable validity conditions.
When a case resumes, retain still-valid information where permitted and repeat affected checks when required. A new document, changed contact detail or expired verification can alter the earlier decision. The resume process should detect those changes instead of treating the old session as a permanent approval.
Also distinguish a draft application from an existing customer or account record. If creation completed but a later step failed, retrying the entire journey can produce duplicates. Preserve application and account references, establish authoritative state and recover the failed task through a controlled process.
Assistance belongs inside the journey
A digital customer may still need help interpreting a document request, resolving a mismatch or understanding a product condition. Assistance can be designed through approved branch, callback, video or other supported routes. The useful measure is whether it completes the intended task while preserving verification and accountability.
Assisted servicing should have appropriate access to the case. A support employee may need the pending step and permitted evidence request without receiving an unrestricted customer file or authority to override an identity decision. Make decision rights visible to staff so escalation reaches the role that can actually resolve the concern.
FATF’s digital-identity guidance examines assurance, governance and the use of digital identity in customer due diligence. It supports considering the reliability of the underlying identity system rather than judging trust by the appearance of the interface. The guidance is international context; it does not replace the rules applicable to an Indian institution. [1]
Give status messages a precise meaning
“Submitted,” “under review,” “created” and “ready to use” should correspond to distinct evidence. The customer needs to know what has completed, what remains and which permitted action comes next. Internal teams need a more detailed task view, linked to the same reference.
If a bank’s approved onboarding route requires a review before operation, an account number should not be presented as proof of immediate usability. Escalate an overdue review to its owner; a service target does not remove the control. Describe pending work truthfully and avoid promising a completion time that the team cannot support.
Customer communication also has boundaries. Explain useful next steps without disclosing protected screening information or sensitive detection rules. Give customers a trusted route to the bank and use approved mechanisms for submitting evidence.
Measure the relationship beyond the first completion screen
Conversion and elapsed onboarding time remain useful. Add time to a usable relationship, evidence-related rework, exception ageing, time to a justified decision, review abandonment, activation defects and successful first intended use. Keep fraud and compliance outcomes alongside those experience measures.
Use consistent denominators. Measuring only applications that completed excludes people who abandoned during review. Counting every account record as an active relationship can hide pending setup. Compare channels and customer contexts carefully, including whether assistance was available and what work it performed.
Link operational outcomes to their causes. A late decision may reflect missing evidence, a specialist queue or an unavailable external service. Each needs a different improvement. A single “drop-off at KYC” label cannot identify which team should act or whether the applicant chose to leave.
Start with one exception and test the whole route
Select a frequent discrepancy. Define the evidence needed, permitted decision paths, access boundaries and accountable owner. Test a legitimate applicant who can resolve it, an applicant whose evidence requires stopping, and a technical failure that interrupts review.
Trace each case through communication, evidence collection, decision, resumption and product setup. Check whether the customer repeats work, whether expired evidence is detected, whether a handoff preserves context and whether the final status matches actual readiness. The test should assess the control and the service together.
Our perspective: make every necessary step useful
Convenience improves when routine customers move promptly and people needing review receive a clear, supported route. Strong verification and a good experience can reinforce one another when the bank designs their operating relationship.
Reduce unnecessary friction. Give every necessary exception an owner, a clear next step and a safe route back into the journey.
That shifts the improvement agenda from counting removed clicks to building dependable beginnings: justified decisions, preserved context, understandable status and a relationship the customer can actually use.
Sources & further reading
- FATF — Guidance on Digital Identity — international guidance on assessing digital identity for customer due diligence.
- RBI — KYC reference edition updated 14 August 2025 — historical background on identification and video identification, not a statement of the current consolidated requirements.
Reviewed on 7 October 2026. Scenarios, calculations, workflow labels and operating proposals are original editorial analysis. They do not establish permitted onboarding routes or exceptions to current requirements. Implementation must use the applicable institution-specific directions and subsequent amendments.
Put this perspective into practice.
Explore the concepts, make a decision and test what changes when the situation changes.
Topics: Digital Onboarding, Perspective, The Banking Balancing Act