KYB onboarding best practices

When a business signs up for your product, every second of friction matters. The difference between a 30-second form and a 3-minute form can cut your conversion rate in half. Yet KYB compliance requires you to collect detailed company information—legal name, registration number, address, legal representatives, beneficial owners. The best onboarding experiences solve this tension by prefilling forms with public data from company registers. Your customer selects their company from a search, and you fill in the rest. Less typing, fewer errors, higher conversion. But register data comes with challenges. Fetching it can be slow, expensive, and unreliable. This guide explains how to navigate these trade-offs to build an onboarding flow that converts.

​

The fundamental trade-off

Company registers are authoritative sources of truth, but they weren’t designed for real-time API access. When you query a register directly, you might wait seconds—or sometimes minutes—for a response. Some registers go down regularly. Pricing varies wildly: a company profile might cost €0.10 in one country and €5 in another. During onboarding, these constraints create a problem. Your prospect is sitting there, waiting. Every second of delay increases the chance they’ll abandon the flow. And since many prospects won’t convert anyway, spending €5 per lookup on everyone who starts your signup form doesn’t make economic sense. The solution is to use different data sources at different stages of your flow: Early in onboarding, when you’re not sure if the prospect will convert, use fast and affordable data. Accept that it might be slightly stale or incomplete—you’ll verify it later. After qualification, when the customer has demonstrated real intent, invest in authoritative verification. Query the register directly, fetch official documents, and cross-check everything. This staged approach gives you the best of both worlds: a fast, high-converting onboarding experience that still meets your compliance requirements.

​

Start with company identification

Before you can prefill anything, you need to know which company your customer represents. This means search. Most onboarding flows already collect the customer’s country early on—to adapt the experience, determine pricing, or define the service area. Use this to scope your search. A country-scoped search is faster and returns more relevant results. When your customer starts typing their company name, you want results to appear immediately. Use streaming search to display results as they arrive from different sources. In countries like Germany, we query both a cached index (fast, ~200ms) and live registers (comprehensive) in parallel. Your customer sees initial results almost instantly, with more appearing as additional sources respond. Search is free and typically returns results in less than 5 seconds. The goal is to make company selection feel effortless.

​

Prefill with onboarding data

Once your customer selects their company, you need to retrieve profile data to populate your forms. This is where the trade-offs become real. If you query the register directly, you might wait 5–30 seconds for a response. In some countries, the register might be down entirely. Your customer is staring at a loading spinner, wondering if something broke. Many will leave. The onboarding endpoint solves this problem. It’s designed specifically for the onboarding use case, with three guarantees:

• Speed: P90 response time under 5 seconds
• Reliability: Over 99% success rate
• Predictability: Fixed €0.50 per request, regardless of country

To achieve this, we make trade-offs. The data might come from a cache rather than a live register query. In countries where register access is expensive, we might use alternative data sources. Some data points that are particularly costly to retrieve may be omitted. These trade-offs are acceptable during onboarding because you’re not making final compliance decisions yet. You’re just trying to reduce friction and collect initial information. The verification comes later, when the customer has proven they’re serious. For a complete list of supported countries, see the Onboarding Profile documentation.

​

Check what documents exist

While you’re fetching onboarding data, also check which official documents are available in the register for this company. This request runs in parallel and is free. Knowing what’s available helps you plan the rest of the flow:

• Documents that exist in the register can be retrieved automatically for verification
• Documents that don’t exist will need to be uploaded by your customer

This avoids the frustrating experience of asking customers to upload documents that you could have fetched yourself.

​

Apply an early filter

With the onboarding data in hand, you can make an initial assessment: can you actually work with this company? Check the basics:

• Is the company active, or has it been dissolved?
• Does the industry or legal form fit your product requirements?
• Are there any obvious red flags in the company structure?

If the answer is no, reject early. This saves everyone time. There’s no point putting a prospect through additional verification steps if you already know you can’t serve them. Early rejection also protects your unit economics. Every step past this point—verification data, document retrieval, manual review—costs money. Don’t spend it on companies that will never become customers.

​

Verify against the register

The data your customer provided—or that you prefilled from the onboarding endpoint—might not match reality. They could be impersonating a different company. They might have provided outdated information. Or the cached data you used for prefilling might have changed since it was last refreshed. At this stage, the customer has demonstrated real intent. They’ve filled out your forms, passed your initial filter, and progressed through your flow. Now it’s worth investing in authoritative data. The verification data endpoint provides comprehensive, verified data:

• Live data: Fetched directly from official government registers
• Completeness: Automatically enriched with information extracted from available documents
• Cross-checking: Data verified against register documents (beta—contact us to enable)

This endpoint is more expensive than the onboarding endpoint, and responses may take longer. That’s the trade-off for authoritative data. But at this stage of your flow, it’s the right trade-off to make. Build this as an asynchronous step. Submit the verification request, then let the customer continue while it processes in the background. Use webhooks to receive the results when they’re ready.

​

Complete with document verification

The final layer of verification is cross-checking against official documents from the register:

• Financial statements reveal the company’s financial health and disclosures
• Register extracts confirm registration details and legal representatives
• Articles of association show company structure and governance rules
• UBO extracts confirm beneficial ownership declarations

Why fetch these from the register rather than asking your customer to upload them? Because documents can be forged, but register responses cannot. When you retrieve a document directly from the source, you know it’s authentic. See our document retrieval guide for details on how document fetching works.

​

Putting it together

The complete flow balances speed with thoroughness: The key insight is that different stages have different requirements. Early on, speed and cost matter most. Later, accuracy and authority matter most. By choosing the right data source for each stage, you get high conversion and strong compliance.

​

The flow

​

Next steps