Customer: DMI Finance
Industry: Lending & Embedded Finance
Products: Tartan Bulk Verification Platform
Primary Use Case: Identity, income, and employment verification at scale

Executive Summary

As DMI Finance scaled its lending business, verification emerged as a critical limiting factor. What had worked at lower volumes - manual uploads, OTP-driven consent, and sequential checks - began to degrade onboarding performance as volumes increased.

At DMI’s scale, even marginal friction in verification translated into material loss in completed applications. Verification performance emerged as a primary determinant of onboarding yield, not a secondary compliance step.

By implementing Tartan’s Bulk Verification platform, DMI restructured how verification fit into its onboarding stack. The result was a system capable of verifying a large volume of users per month, while reducing onboarding drop-offs, improving operational efficiency, and preserving regulatory compliance - all without changes to their existing Loan Origination System.

Reducing drop-offs materially shifted downstream conversion economics, not just user experience metrics. The ability to verify users at scale became directly correlated with revenue realization speed.

Background: When Growth Exposes Fragility

DMI operates in a high-velocity lending environment where onboarding speed, data accuracy, and compliance are tightly coupled to conversion and risk outcomes.

At a moderate scale, their verification workflows were manageable. But as user volumes increased, the architecture began to show strain.

Processes that perform adequately at low volume tend to fail non-linearly as scale increases. Operational workarounds masked systemic inefficiencies until volume made them unavoidable.

Verification depended on manual CSV uploads handled by internal teams, OTP-based consent flows that required synchronous user participation, and separate processes for PAN, UAN, and employment checks. Each step introduced delay. Each delay increased the likelihood of drop-off.

Growth exposed architectural limits that could not be solved through incremental optimization.

The Core Problem: Verification Was Tied to User Patience

The biggest constraint wasn’t data availability - it was flow dependency.

Verification success rates were constrained less by data availability and more by real-time user behavior. Each additional synchronous step increased abandonment probability disproportionately.

User-driven verification flows created variance that propagated into downstream credit decision timelines. At scale, verification latency became indistinguishable from conversion loss.

Insight:
In high-volume onboarding systems, user patience becomes an unreliable dependency. Any workflow that assumes synchronous user cooperation will eventually cap conversion - not because users fail verification, but because they disengage before it completes.

What DMI Needed to Change

DMI was not looking for another point solution. The requirement was architectural.

The objective was not faster verification in isolation, but predictable verification under load. DMI required verification capacity that scaled independently of staffing, retries, or user intervention.

The system had to remain compliant by design, not through enforcement after failure. Any solution that required core LOS changes would have introduced unacceptable execution risk.

Insight:
At scale, the cost of modifying core systems often exceeds the cost of verification inefficiency itself. DMI’s priority was to isolate verification evolution from core lending logic, preserving long-term optionality.

The Tartan Implementation

DMI implemented Tartan’s Bulk Verification platform as a standalone verification layer.

Instead of triggering verifications one user at a time, DMI teams could submit large batches of records - via CSV or API - and receive verified outputs asynchronously. PAN, UAN, and employment checks were handled in a single consolidated workflow.

Bulk execution fundamentally altered the unit economics of verification. Asynchronous processing removed verification from the critical path of user interaction.

Crucially, Tartan’s platform preserved consent logic within bulk operations, ensuring regulatory compliance without forcing users through repeated OTP-driven flows. Consent logic embedded at the system level eliminated the need for repeated user-side remediation.

Decoupling verification from the LOS preserved long-term architectural flexibility.

Operational Impact: From Manual Effort to Systemic Scale

Before Tartan, verification throughput was constrained by manual intervention, user responsiveness, and batch processing limitations.

After implementation, verification became asynchronous, bulk-driven, and operationally lightweight.

Verification throughput was no longer bounded by manual coordination or batch timing. Operational predictability improved as verification volume increased, while exception handling decreased as a percentage of total verifications.

Operational effort shifted from remediation to oversight.

Insight:
True scale is not measured by peak throughput, but by how little operational effort is required to sustain it. DMI’s verification system became quieter as it grew - a signal of structural efficiency.

Business Impact: Measurable Gains in Onboarding Performance

With verification friction removed from critical onboarding paths, DMI observed:

• A large volume of users verified consistently per month.

• A significant reduction in onboarding drop-offs, driven by fewer stalled or failed verification steps.

Drop-off reduction at this stage of onboarding disproportionately improved funded application rates. Verification ceased to be a source of silent revenue leakage.

Higher verification reliability reduced variance in conversion forecasting. Faster progression through onboarding improved time-to-revenue without changing credit policy.

Insight:
When verification reliability increases, forecasting confidence improves. This has second-order effects - from staffing plans to capital deployment - that extend well beyond onboarding metrics.

Why This Worked

The success of the implementation came down to one principle: verification was treated as infrastructure, not interaction.

The system succeeded because it removed unnecessary coupling between users, operations, and verification execution. Scale was achieved through architectural simplification, not process complexity.

Verification reliability improved by reducing dependencies, not adding controls. The model favored systemic correctness over edge-case recovery.

Most verification systems fail gracefully at low scale and catastrophically at high scale. DMI avoided this by designing for peak load and operational indifference - where success does not depend on human vigilance, retries, or perfect user behavior.

Customer Perspective

“As our onboarding volumes increased, verification became one of the biggest constraints on growth. Tartan allowed us to scale verification in a way that was operationally sound, compliant, and far more resilient than what we had before.”
- Chief Product Officer, DMI Finance

What This Enables Going Forward

With a scalable verification layer in place, DMI is now positioned to support significantly higher onboarding volumes without increased drop-off risk.

Verification infrastructure decisions now have multi-year strategic consequences. What appears as an onboarding optimization can become a growth constraint at scale.

DMI’s experience reinforced that verification must be designed for peak volume, not average load. Stability at scale became a competitive advantage rather than a maintenance cost.

About Tartan

Tartan provides verification infrastructure for financial institutions - covering identity, income, and employment - designed to operate at scale, reduce friction, and improve conversion across onboarding, underwriting, and compliance workflows.