FinTech Focus

KYB Groundhog Day

I was helping a partner bank spin up a new SMB lending flow. Day one looked like this:

1. Paste company name into Companies House → download PDF.

2. Manually cross-check address lines.

3. Screenshot the website in case auditors ask “how did you know what they do?”.

4. Run four separate AML screens, most of which came back as false positives.

💀 38 minutes later, I still hadn’t started the credit logic.

Multiply that by a few hundred sign-ups per day, and you’ve burned entire sprints on clerical work that adds zero product value—and still miss subtle fraud signals.

Every KYB-heavy fintech team I know is living the same loop.

Diligent flips that loop on its head—hand the inputs to an AI agent that follows your playbook, returns a structured risk report, and only escalates edge cases.

AI agents for fintech risk & compliance

Diligent is basically “Copilot for KYB analysts.” Non-technical friend version: “Paste a business, get an instant red-flag report instead of doing six browser tabs of detective work.”

The magic: the agent doesn’t stop at registry data—it spiders the web, scores website reputation, ingests customer-uploaded docs, and explains every step, so compliance can sign off. 

Today it’s live with a handful of European neobanks and payment processors (the founders are YC W24 / S23, a 2-person London crew with deep BNPL & fraud-risk scars). 

Key Information

🧩 Full-stack KYB: CDD, doc review, sanctions/PEP alert triage—switch on/off per workflow.

💰 Funded during YC S23 batch with a $500k pre-seed round led by Y Combinator and fintech-angel operators; seed raise now in motion.

🏗️ LLM-explainability baked in every decision ships with an audit-ready rationale string; no more “the model said so”.

The Bottom Line

Diligent does one thing remarkably well: turn raw business inputs into an explainable risk decision in seconds. At a mid-four-figure monthly bill, it pays for itself the first time you cut a full-time analyst or ship that dreaded remediation project ahead of audit.

Bigger picture: AI agents are finally good enough to own repetitive compliance ops—freeing devs to build features customers notice. If you’re drowning in KYB tickets, give the sandbox a spin before you write another Selenium script.

Try it: https://www.godiligent.ai/

Help Me Improve FinTex