On this page

Yes, a no-code route exists for bank compliance teams to build and change onboarding journeys without waiting on IT. This case study walks through a representative bank scenario: what the dependency looks like, what compliance owns after the change, and which results Duna has reported. It is a scenario built from how Duna customers configure journeys, not a single named bank. Outcome figures come from Duna's reported customer results, and the quotes section describes what each team should be able to confirm rather than quoting individuals.

The situation: every policy change became an engineering ticket

Consider a mid-sized bank onboarding business customers across several countries. Compliance owns the KYB and KYC policy: which documents are required, when beneficial ownership must be verified, and which risk scores trigger enhanced review. Engineering owns the onboarding application where those rules live.

The result is a queue. A compliance officer decides that a high-risk jurisdiction needs an extra document. The request becomes a ticket, waits for sprint planning, gets built, tested, and released. Ownership is also unclear. Compliance defines the rule but cannot see how it was implemented, and engineering implements a rule it does not fully own.

The stakes are practical:

  • Policy drift: production can lag behind the written policy for weeks.

  • Applicant friction: static forms ask every applicant for the same data, regardless of risk, which lowers conversion.

  • Analyst workload: reviewers compensate manually for gaps the journey does not handle.

  • Audit exposure: auditors ask when a rule changed and who approved it, and the answer sits across ticket systems and release notes.

Manual review is costly at scale. ioNova AI (May 12, 2026) estimates a direct cost of $15 to $25 per alert, and says that where 99% of alerts are false positives, the annual cost of manual reviews can reach $370,000 to $620,000 in high-volume environments.

What compliance changed once it owned the journey

The fix was to move journey logic out of application code and into a configuration layer that compliance operates directly. In Duna's model, Duna Onboard provides no-code onboarding journeys with pre-built modules, and the policy engine holds the compliance logic.

The first changes a compliance team would typically make in production:

  • Reorder steps so that document collection depends on the applicant's entity type and country.

  • Add a verification checkpoint for a jurisdiction under heightened scrutiny.

  • Adjust risk thresholds so low-risk applicants move through without manual review.

  • Change reviewer routing, for example sending high-risk cases to a four-eye review.

The policy engine translates these rules into code, so compliance defines the compliant state and the system executes it. Changes can be back-tested before release. Engineering still maintains integrations and infrastructure, but no longer touches policy logic, field requirements, or routing rules. Our guide on how a policy engine keeps compliance current explains the mechanics in more detail.

Results Duna has reported

The three outcome areas that matter are IT dependency, onboarding performance, and analyst efficiency. Duna reports the following across its customer base, including Plaid, Moss, and SeQura:

Area

Reported result

Onboarding speed

10.6x faster onboarding

Conversion

37% increase

Analyst efficiency

4.8x

These are aggregate figures, not results from one bank. Neurons Lab (March 24, 2026) points the same way, stating that AI for compliance can cut KYC processing time from days to hours and free officers for complex cases.

The IT-dependency metric differs by institution, so a bank should measure its own baseline before any pilot. Three numbers capture it:

  1. Engineering tickets raised for policy or journey changes per quarter.

  2. Median days from an approved compliance decision to production.

  3. Share of journey changes released without an engineering deployment.

Duna Onboard is built to go live within days, and customer-specific figures for these three measures should come from the pilot, not from vendor claims.

Governance: self-serve configuration with audit control

Removing engineering from the change path does not remove control. Three mechanisms keep it intact.

Role permissions. Not every compliance user can publish to production. Drafting, reviewing, and approving are separate roles.

Versioned policy logic. Each change is stored as a version, so a bank can show which rules applied to a given applicant on a given date.

Audit trails. Duna logs interactions, decisions, and policy changes in one place. An auditor's question about who changed a threshold and when has a single answer.

Teams starting this way should standardize templates for common journeys first, such as domestic SME and cross-border corporate, and test changes against historical cases before release. Some controls should stay behind formal approval even in a self-serve model: changes to sanctions screening scope, risk-score weightings, and anything that alters the bank's documented risk appetite.

What each team should be able to confirm

A pilot succeeds when both sides can state the outcome plainly. Compliance leadership should be able to say that approved policy changes reach production without an engineering release. Engineering or IT governance should be able to say that policy and journey tickets have dropped, and that integrations and security remain under its control. If either statement cannot be made with a number behind it, the pilot has not finished.

Next step

The change comes down to one sentence: policy decisions moved from a ticket queue into a governed configuration layer that compliance operates and engineering oversees.

To test it, run a pilot on one onboarding journey, record the three IT-dependency baselines above, and have compliance make the next three policy changes itself. Banks evaluating this approach can review Duna's banking solution and compare architectures in the KYC and KYB orchestration platform guide.

Yes, a no-code route exists for bank compliance teams to build and change onboarding journeys without waiting on IT. This case study walks through a representative bank scenario: what the dependency looks like, what compliance owns after the change, and which results Duna has reported. It is a scenario built from how Duna customers configure journeys, not a single named bank. Outcome figures come from Duna's reported customer results, and the quotes section describes what each team should be able to confirm rather than quoting individuals.

The situation: every policy change became an engineering ticket

Consider a mid-sized bank onboarding business customers across several countries. Compliance owns the KYB and KYC policy: which documents are required, when beneficial ownership must be verified, and which risk scores trigger enhanced review. Engineering owns the onboarding application where those rules live.

The result is a queue. A compliance officer decides that a high-risk jurisdiction needs an extra document. The request becomes a ticket, waits for sprint planning, gets built, tested, and released. Ownership is also unclear. Compliance defines the rule but cannot see how it was implemented, and engineering implements a rule it does not fully own.

The stakes are practical:

  • Policy drift: production can lag behind the written policy for weeks.

  • Applicant friction: static forms ask every applicant for the same data, regardless of risk, which lowers conversion.

  • Analyst workload: reviewers compensate manually for gaps the journey does not handle.

  • Audit exposure: auditors ask when a rule changed and who approved it, and the answer sits across ticket systems and release notes.

Manual review is costly at scale. ioNova AI (May 12, 2026) estimates a direct cost of $15 to $25 per alert, and says that where 99% of alerts are false positives, the annual cost of manual reviews can reach $370,000 to $620,000 in high-volume environments.

What compliance changed once it owned the journey

The fix was to move journey logic out of application code and into a configuration layer that compliance operates directly. In Duna's model, Duna Onboard provides no-code onboarding journeys with pre-built modules, and the policy engine holds the compliance logic.

The first changes a compliance team would typically make in production:

  • Reorder steps so that document collection depends on the applicant's entity type and country.

  • Add a verification checkpoint for a jurisdiction under heightened scrutiny.

  • Adjust risk thresholds so low-risk applicants move through without manual review.

  • Change reviewer routing, for example sending high-risk cases to a four-eye review.

The policy engine translates these rules into code, so compliance defines the compliant state and the system executes it. Changes can be back-tested before release. Engineering still maintains integrations and infrastructure, but no longer touches policy logic, field requirements, or routing rules. Our guide on how a policy engine keeps compliance current explains the mechanics in more detail.

Results Duna has reported

The three outcome areas that matter are IT dependency, onboarding performance, and analyst efficiency. Duna reports the following across its customer base, including Plaid, Moss, and SeQura:

Area

Reported result

Onboarding speed

10.6x faster onboarding

Conversion

37% increase

Analyst efficiency

4.8x

These are aggregate figures, not results from one bank. Neurons Lab (March 24, 2026) points the same way, stating that AI for compliance can cut KYC processing time from days to hours and free officers for complex cases.

The IT-dependency metric differs by institution, so a bank should measure its own baseline before any pilot. Three numbers capture it:

  1. Engineering tickets raised for policy or journey changes per quarter.

  2. Median days from an approved compliance decision to production.

  3. Share of journey changes released without an engineering deployment.

Duna Onboard is built to go live within days, and customer-specific figures for these three measures should come from the pilot, not from vendor claims.

Governance: self-serve configuration with audit control

Removing engineering from the change path does not remove control. Three mechanisms keep it intact.

Role permissions. Not every compliance user can publish to production. Drafting, reviewing, and approving are separate roles.

Versioned policy logic. Each change is stored as a version, so a bank can show which rules applied to a given applicant on a given date.

Audit trails. Duna logs interactions, decisions, and policy changes in one place. An auditor's question about who changed a threshold and when has a single answer.

Teams starting this way should standardize templates for common journeys first, such as domestic SME and cross-border corporate, and test changes against historical cases before release. Some controls should stay behind formal approval even in a self-serve model: changes to sanctions screening scope, risk-score weightings, and anything that alters the bank's documented risk appetite.

What each team should be able to confirm

A pilot succeeds when both sides can state the outcome plainly. Compliance leadership should be able to say that approved policy changes reach production without an engineering release. Engineering or IT governance should be able to say that policy and journey tickets have dropped, and that integrations and security remain under its control. If either statement cannot be made with a number behind it, the pilot has not finished.

Next step

The change comes down to one sentence: policy decisions moved from a ticket queue into a governed configuration layer that compliance operates and engineering oversees.

To test it, run a pilot on one onboarding journey, record the three IT-dependency baselines above, and have compliance make the next three policy changes itself. Banks evaluating this approach can review Duna's banking solution and compare architectures in the KYC and KYB orchestration platform guide.