What's the most cost-effective way to run wallet screening using an AML screening tool at scale?

The most cost-effective way is usually to screen only where the risk changes, not every time a wallet is touched.

Practical approach

1. Screen at onboarding / first contact

   • New wallet address
   • New counterparty
   • New customer profile

2. Use event-driven rescreening Rescreen only when one of these changes:

   • wallet receives funds from a new source
   • wallet interacts with a high-risk entity
   • a transaction exceeds a threshold
   • sanctions/PEP/adverse-risk lists update
   • the wallet’s behavior changes materially

3. Tier your screening depth

   • Low risk: lightweight chain analytics + sanctions check
   • Medium risk: deeper graph lookback, entity clustering, exposure scoring
   • High risk: full AML review, source-of-funds, manual escalation

4. Cache results aggressively

   • Store screening scores, timestamps, and decision outcomes
   • Reuse results for the same wallet until a trigger invalidates them
   • Deduplicate repeated addresses/entities across users

5. Batch where latency isn’t critical

   • Run periodic bulk screening jobs for dormant wallets
   • Process queues off-peak
   • Use async workflows for non-blocking compliance checks

6. Pre-filter before vendor calls

   • Basic rules engine first
   • Only send wallets that pass thresholds into the expensive AML API
   • This can cut API volume significantly

7. Use risk-based thresholds

   • Avoid treating every wallet equally
   • Focus resources on:
     • high-value wallets
     • mixers
     • cross-chain bridges
     • darknet exposure
     • sanctioned jurisdiction links

8. Negotiate pricing model carefully

   • If possible, prefer:
     • per-screening volume discounts
     • monthly committed tiers
     • cached/API-reuse support
     • webhook/list-update pricing instead of per-call rescreening

Best operating model

For scale, the cheapest reliable pattern is:

Rules engine + cached wallet profiles + event-triggered rescreening + periodic bulk refresh + manual review only for high-risk cases.

What to avoid

• Screening every transaction blindly
• Re-screening identical wallets repeatedly
• Using full-depth AML scoring for all addresses
• Calling the vendor API without deduplication or caching

If you want, I can also outline a reference architecture or cost model for wallet screening at 10k / 1M / 100M addresses.