Merchant Onboarding — PNC
How a seven-stage screening pipeline cut underwriting time 38% while blocking 27% more fraud.
An enterprise screening dashboard covering the full seven-stage merchant onboarding pipeline — from pre-screening to web-content analysis — with a GenAI layer that surfaces risk in plain language and keeps payment environments secure and compliant.
Seven-stage screening process
Each stage compounds confidence in a merchant's legitimacy. The dashboard maps one-to-one onto this flow so every drop-off is traceable to a real stage.
Pre-screening
Businesses evaluate potential PSPs or gateways that align with their needs.
Identity checks
Rigorous identity verification ensures the business's legitimacy and minimizes fraud risks.
Underwriting
Assess the business's financial stability, creditworthiness, and risk profile.
Merchant history screening
Examine past performance; prevalent issues help gauge reliability and trustworthiness.
Business & operational analysis
Understand the business model, operations, and scalability to tailor a payment solution.
Compliance setup
Compliance with regulations such as anti-money-laundering and data protection is non-negotiable.
Web content analysis
The business's online presence is scrutinized, including website content and security protocols.
Explore the live dashboard
A working slice of the product. Switch the time range to watch the screening funnel, KPIs, live risk signals, and the GenAI narrative respond in real time. Hover any metric for a plain-language definition.
Merchant Onboarding — PNC · Live Demo
Seven-stage screening pipeline · as of May 2025
Screening Pipeline Funnel
7D · merchants per stageLive Risk Signals
from screening engine- Medium
Identity mismatch
Document vs. registry conflicts ↑
- High
High-risk merchant category
MCC flagged for enhanced review
- Low
AML / sanctions matches
17 cases flagged for review
- Medium
Adverse merchant history
Prior chargebacks detected
- Low
Website / content risk
Insecure checkout on 4 sites
AI Screening Narrative
GenAIOver the last 7D, 14,210 merchants entered pre-screening and 4,980 cleared web-content analysis (35% end-to-end). The largest drop-off is "Identity Checks" → "Underwriting" at 19%, driven by identity mismatches and adverse merchant history in high-risk categories. Recommended action: trigger the GenAI document-assist and adverse-media screening earlier, and fast-track low-risk merchants to protect approval velocity without loosening AML controls.
Diverse payment methods
Accept cards, digital wallets, and NetBanking so customers pay with their preferred method.
Speed & simplicity
Fast, efficient payment flows streamline transactions for both customers and merchants.
Security & trust
Robust, encrypted payment platforms give merchants confidence in every transaction.
Transaction monitoring
Real-time insights reveal sales patterns and peak hours to optimize operations.
Metrics glossary
plain-language definitions for every metric on this dashboard- Onboarding Approval Rate
- Share of pre-screened merchants that clear all seven stages and are approved to transact.
- Identity Pass Rate
- Percentage of merchants that clear identity verification on first pass.
- Fraud Blocked
- Count of applications stopped by identity, AML, or history checks before approval.
- Avg. Underwriting Time
- Average time to assess financial stability, creditworthiness, and risk profile.
- PSP / Gateway
- Payment Service Provider / payment gateway — the platform that processes a merchant's card and digital payments.
- Underwriting
- Risk assessment of a merchant's finances, credit, and business model to decide approval terms.
- AML / Sanctions
- Anti-Money-Laundering and global sanctions screening to prevent onboarding prohibited entities.
- MCC
- Merchant Category Code — classifies a business type; some categories carry higher inherent risk.
- Chargeback
- A forced payment reversal initiated by a cardholder's bank; high rates signal merchant risk.
- pp (percentage points)
- The absolute difference between two percentages. +4pp means a rate rose from, e.g., 30% to 34%.
The Challenge
Merchant screening ran through seven mandated stages — pre-screening, identity checks, underwriting, merchant-history screening, business and operational analysis, compliance setup, and web-content analysis. Each stage had an owner, a tool, and its own evidence trail, and none of them shared a view. Legitimate merchants waited behind manual checks while genuinely risky ones were spotted late.
Why It Mattered
Every stage compounds confidence in a merchant's legitimacy: identity verification minimises fraud, underwriting quantifies financial risk, history screening surfaces past issues, and web-content analysis validates the live online presence and security posture. Weak visibility at any stage becomes a chargeback, a fine, or a fraud loss downstream.
The Strategy
One dashboard mapped one-to-one onto the seven stages, so every drop-off is traceable to a real control rather than a black box. A GenAI layer accelerates the pipeline without loosening it: it extracts and validates identity documents, runs adverse-media and sanctions screening, summarises merchant history, and reads the live funnel to write a plain-language risk narrative with ranked, explainable actions — each linked back to the signal behind it.
The Impact
Underwriting time fell 38% and 27% more fraudulent applications were blocked, at a 35% end-to-end approval rate across all seven stages. Merchants and sub-merchants gained faster access to diverse payment methods — cards, digital wallets, NetBanking — with encrypted transaction security and real-time monitoring that reveals sales patterns and peak hours.
What I Carry Forward
Compliance stages are not overhead to hide; made visible, they become the product's credibility. Speed came from removing ambiguity between stages, not from removing controls.
Behind the work
My scope, the guardrails I worked inside, what the research changed, and how the work travelled from sketch to production.
My role & scope
Product design lead across all seven screening stages: stakeholder mapping with risk, compliance and underwriting owners, the unified funnel IA, the GenAI risk-narrative layer, and the component states for each stage. Paired with 5 engineers and a compliance SME.
Constraints & guardrails
- AML, KYC and data-protection rules fixed the stage order — the flow could be clarified, never shortened.
- Seven legacy tools with different data models, integrated behind one view without replacing any of them.
- 16-week delivery window tied to an acquiring-portfolio launch.
- Every automated decision had to remain auditable and reversible by a human reviewer.
What research changed
- Process mapping with 9 stage owners found the same merchant evidence re-requested in 3 separate stages.
- Queue analytics: 41% of total onboarding time sat in hand-off waits, not in the checks themselves.
- Reviewers said the blocker was not the decision but reconstructing what earlier stages had already found.
The pivot
An early concept collapsed the seven stages into three simplified ones. Compliance rejected it — the mandated stages must stay visible for audit. We kept all seven and instead removed ambiguity between them with a shared evidence trail and a plain-language narrative.
UX metric → business impact
Service blueprint of the seven stages with owners, tools and hand-off waits marked in red.
Interactive funnel prototype with reviewer-decision states and accessibility annotations for the queue tables.
Live screening dashboard: stage funnel, KPI band, ranked risk signals and the GenAI narrative with linked evidence.
Retrospective
Run a moderated study with sub-merchant applicants, not just internal reviewers — the outside-in view of the wait would have sharpened the pre-screening stage.
Compliance stages are not overhead to hide. Made visible and legible, they become the product's credibility — and the speed comes from removing ambiguity between stages, not controls.
“The screening story used to live in seven different places. Putting it in one narrative changed how fast we could say yes — and how confidently we could say no.”