How Duna integrates with banks’ existing systems
Hande Kahraman

Hande Kahraman

On this page

Banks have a lot to keep up with when it comes to verifying the identity of business customers. AI agents can now pass as real people, making fraudulent applications harder to detect. Anti-money laundering requirements keep changing. Periodic reviews and customer updates also add to compliance teams’ workloads long after initial onboarding.

Financial institutions are adopting new technology to meet these demands. The first question they ask about any new compliance system is how it connects to what they already run, and whether it replaces systems they’ve spent years building. Duna works on top of those systems. It reads from them, runs the checks, and writes each decision back.

Duna’s policy engine keeps information and decisions in sync, allowing banks to combine their existing systems with the Duna components they choose.

The policy engine makes integration easier

The policy engine is the foundation of every Duna integration. It’s responsible for:

  • Applying policies. The bank defines its policies and works with Duna to prepare them for automation. This involves turning policies into clear instructions that can be translated into code. The policy engine then applies those instructions using deterministic rules to produce predictable, repeatable outcomes.

  • Coordinating systems. Think of a spider in the middle of a web. Every update or new piece of evidence flows into the policy engine, which then directs information and actions wherever they’re needed.

The policy engine can coordinate a few components or manage policies and decisions across the technology environment. Because the engine manages information and decisions centrally, banks can add or replace individual technologies without rebuilding connections between every system.

Integrating Duna with existing technology

With the policy engine at the center, each institution can decide what to keep in-house and what to adopt from Duna for onboarding, data enrichment, case management, and CRM or system of record.

Diagram of Duna’s policy engine at the center of onboarding experience, data platform, case management, and a bank’s internal systems

Onboarding UI

Banks can keep their own onboarding interface and send the information they collect to Duna through an API. The policy engine determines which checks have been completed and which are still needed.

Or they can choose Duna’s white-label UI to collect customer information directly. This could be a better option if their current interface is hard to maintain or creates friction during onboarding.

Data enrichment

There are three options for gathering the external evidence needed to verify customers:

  • Manage provider relationships and technical integrations in-house.

  • Keep the data-provider contracts and have Duna handle the technical integrations.

  • Let Duna manage data enrichment end to end, including provider contracts, integrations, and data quality.

Case management

If a bank keeps its own case management system, Duna provides the collected evidence and identifies its source, while the bank’s reviewers handle case evaluations in that system.

Banks that want to automate more of this work can adopt Duna’s case management capabilities, which use AI agents to evaluate evidence and recommend decisions based on the bank’s policies. Analysts see each step an agent took and the evidence behind it, and can change the outcome. The change is recorded with the case. This could be a better option when the bank’s current workflow tools require too many manual steps or have permission structures that limit what these agents can do.

CRM / system of record

Financial institutions can keep their CRM, with Duna synchronizing updated customer information and decisions back to that system. Or they can keep the customer record in Duna.

What the bank’s team does, and what Duna does

Duna’s deployment team builds the integrations, configures the bank’s policies in the policy engine, and sets up the onboarding flow. The bank’s teams provide what only they can give. Compliance joins the policy workshops, IT joins a technical scoping session and the security review, and analysts are trained on the review screen.

A first project can run on a defined set of cases, with data exchanged by file, before any system is connected. The bank connects its systems once the results justify it.

How data flows between connected systems

The following examples show how Duna’s policy engine coordinates data and actions across systems during initial onboarding, periodic reviews, and reboarding.

Initial onboarding

  • Collect customer information. Data enters through the institution’s onboarding UI or Duna’s. The policy engine checks what’s already available and what’s missing.

  • Gather external evidence. The engine directs AI agents to gather registry, fraud, and sanctions information through the data enrichment component. If any evidence is missing, the engine can instruct an agent to try again or request it from the customer through the onboarding UI.

  • Assess the customer. The policy engine automatically approves straightforward cases and sends those requiring further assessment to case management.

  • Update the customer record. Once the decision is made, the policy engine ensures it’s sent to the bank’s CRM or system of record, or kept in Duna.

Periodic reviews and reboarding

Periodic reviews could require analysts to revisit earlier assessments, while changes in company ownership or a customer’s request for a new banking product can trigger reboarding.

New evidence can come from any connected system, from the CRM to a data provider. The policy engine receives the information, applies the bank’s policies, and directs AI agents to request the necessary data from the relevant system.

In a review, the engine first checks which evidence on file is still valid. It then gathers what has changed from the data providers and asks the customer only for what is still missing.

Choose what to keep and where to invest

Duna’s AI assessment maps the systems an analyst uses on a case, checks how ready the policies are for automation, and estimates the return for each process.

When deciding whether to use their own systems or Duna’s, compliance and operations leaders should ask themselves:

  • Where are we willing to make a change?

  • Where are we commercially locked in?

  • Is the ROI worth the change?

One organization we worked with onboards customers through direct and indirect channels. They started out wanting to optimize their partner channel, but after seeing how much friction was involved in direct onboarding, they saw more potential for ROI in that process.

They decided to run a proof of concept focused on the customer experience, with Duna taking responsibility for the onboarding interface and data collection.

Another institution believed their biggest opportunity was around efficiency in compliance workflows. But the ROI analysis showed that improving conversion offered higher returns than reducing compliance costs.

The best opportunity for ROI could come from a different part of the onboarding process than leaders initially expect.

Build flexibility into compliance operations

Financial institutions need a way to manage compliance processes across their systems as they respond to new fraud tactics, regulations, and technologies. Duna’s policy engine allows banks to adapt without disrupting the systems they rely on.


Book a Duna AI assessment to find where to start. Book an AI assessment.

Banks have a lot to keep up with when it comes to verifying the identity of business customers. AI agents can now pass as real people, making fraudulent applications harder to detect. Anti-money laundering requirements keep changing. Periodic reviews and customer updates also add to compliance teams’ workloads long after initial onboarding.

Financial institutions are adopting new technology to meet these demands. The first question they ask about any new compliance system is how it connects to what they already run, and whether it replaces systems they’ve spent years building. Duna works on top of those systems. It reads from them, runs the checks, and writes each decision back.

Duna’s policy engine keeps information and decisions in sync, allowing banks to combine their existing systems with the Duna components they choose.

The policy engine makes integration easier

The policy engine is the foundation of every Duna integration. It’s responsible for:

  • Applying policies. The bank defines its policies and works with Duna to prepare them for automation. This involves turning policies into clear instructions that can be translated into code. The policy engine then applies those instructions using deterministic rules to produce predictable, repeatable outcomes.

  • Coordinating systems. Think of a spider in the middle of a web. Every update or new piece of evidence flows into the policy engine, which then directs information and actions wherever they’re needed.

The policy engine can coordinate a few components or manage policies and decisions across the technology environment. Because the engine manages information and decisions centrally, banks can add or replace individual technologies without rebuilding connections between every system.

Integrating Duna with existing technology

With the policy engine at the center, each institution can decide what to keep in-house and what to adopt from Duna for onboarding, data enrichment, case management, and CRM or system of record.

Diagram of Duna’s policy engine at the center of onboarding experience, data platform, case management, and a bank’s internal systems

Onboarding UI

Banks can keep their own onboarding interface and send the information they collect to Duna through an API. The policy engine determines which checks have been completed and which are still needed.

Or they can choose Duna’s white-label UI to collect customer information directly. This could be a better option if their current interface is hard to maintain or creates friction during onboarding.

Data enrichment

There are three options for gathering the external evidence needed to verify customers:

  • Manage provider relationships and technical integrations in-house.

  • Keep the data-provider contracts and have Duna handle the technical integrations.

  • Let Duna manage data enrichment end to end, including provider contracts, integrations, and data quality.

Case management

If a bank keeps its own case management system, Duna provides the collected evidence and identifies its source, while the bank’s reviewers handle case evaluations in that system.

Banks that want to automate more of this work can adopt Duna’s case management capabilities, which use AI agents to evaluate evidence and recommend decisions based on the bank’s policies. Analysts see each step an agent took and the evidence behind it, and can change the outcome. The change is recorded with the case. This could be a better option when the bank’s current workflow tools require too many manual steps or have permission structures that limit what these agents can do.

CRM / system of record

Financial institutions can keep their CRM, with Duna synchronizing updated customer information and decisions back to that system. Or they can keep the customer record in Duna.

What the bank’s team does, and what Duna does

Duna’s deployment team builds the integrations, configures the bank’s policies in the policy engine, and sets up the onboarding flow. The bank’s teams provide what only they can give. Compliance joins the policy workshops, IT joins a technical scoping session and the security review, and analysts are trained on the review screen.

A first project can run on a defined set of cases, with data exchanged by file, before any system is connected. The bank connects its systems once the results justify it.

How data flows between connected systems

The following examples show how Duna’s policy engine coordinates data and actions across systems during initial onboarding, periodic reviews, and reboarding.

Initial onboarding

  • Collect customer information. Data enters through the institution’s onboarding UI or Duna’s. The policy engine checks what’s already available and what’s missing.

  • Gather external evidence. The engine directs AI agents to gather registry, fraud, and sanctions information through the data enrichment component. If any evidence is missing, the engine can instruct an agent to try again or request it from the customer through the onboarding UI.

  • Assess the customer. The policy engine automatically approves straightforward cases and sends those requiring further assessment to case management.

  • Update the customer record. Once the decision is made, the policy engine ensures it’s sent to the bank’s CRM or system of record, or kept in Duna.

Periodic reviews and reboarding

Periodic reviews could require analysts to revisit earlier assessments, while changes in company ownership or a customer’s request for a new banking product can trigger reboarding.

New evidence can come from any connected system, from the CRM to a data provider. The policy engine receives the information, applies the bank’s policies, and directs AI agents to request the necessary data from the relevant system.

In a review, the engine first checks which evidence on file is still valid. It then gathers what has changed from the data providers and asks the customer only for what is still missing.

Choose what to keep and where to invest

Duna’s AI assessment maps the systems an analyst uses on a case, checks how ready the policies are for automation, and estimates the return for each process.

When deciding whether to use their own systems or Duna’s, compliance and operations leaders should ask themselves:

  • Where are we willing to make a change?

  • Where are we commercially locked in?

  • Is the ROI worth the change?

One organization we worked with onboards customers through direct and indirect channels. They started out wanting to optimize their partner channel, but after seeing how much friction was involved in direct onboarding, they saw more potential for ROI in that process.

They decided to run a proof of concept focused on the customer experience, with Duna taking responsibility for the onboarding interface and data collection.

Another institution believed their biggest opportunity was around efficiency in compliance workflows. But the ROI analysis showed that improving conversion offered higher returns than reducing compliance costs.

The best opportunity for ROI could come from a different part of the onboarding process than leaders initially expect.

Build flexibility into compliance operations

Financial institutions need a way to manage compliance processes across their systems as they respond to new fraud tactics, regulations, and technologies. Duna’s policy engine allows banks to adapt without disrupting the systems they rely on.


Book a Duna AI assessment to find where to start. Book an AI assessment.