Footprint
Identity infrastructure and AI investigation workflows
Footprint combines person and business onboarding, watchlist checks, bank linking, a data vault, and review operations. Percy adds agents that investigate cases and feed configured workflows. Trust Fabric records evidence and shared investigation context. Some checks use named providers. Its transaction-monitoring investigation claims do not establish a standalone AML detection engine, payment authorization control, or regulatory filing service.
Our assessment
A combined onboarding and investigation platform with detailed public integration docs. Its connected checks and bounded agent actions need a product-specific design; broad branding does not establish every financial-crime control.
Best fit and limitations
Useful when a team wants onboarding, verification evidence, sensitive-data handling, and review workflows in one integration. Percy also merits evaluation as an investigation layer beside existing screening and monitoring systems. Confirm which detection systems and data contracts remain outside Footprint.
- Transaction-alert investigation and bank-transaction retrieval do not establish a standalone AML scenario engine or payment-fraud authorization service.
- KYB and synthetic-fraud checks use named providers. Confirm their coverage, data rights, billing, and failure behavior.
- The public POST /onboardings API is marked Preview. Some APIs vary by account or need enablement; a published guide is not proof of your entitlement.
- A KYB flow that verifies beneficial owners cannot complete through the headless onboarding call alone; the integration guide directs those users to an onboarding session.
- No reviewed source establishes credit underwriting, merchant reserves, settlement execution, or direct regulatory submission. Marketing country totals are not a product-by-country availability table.
Tools 11
KYC and identity-document verificationPerson onboarding+
- Collect identity data, government documents, and selfies in hosted or embedded flows
- Support capture, liveness, and device handoff within the document flow
KYB and beneficial-owner onboardingConnected business verification+
- Run business or EIN checks through Middesk or Baselayer
- Combine business onboarding with separate beneficial-owner identity checks
AML watchlist screeningCustomer screening+
- Check OFAC, political exposure, and adverse-media lists with exact or fuzzy matching
- Configure ongoing monthly rescreening; check-completion events support follow-up
PlaybooksCollection, checks, and policy decisions+
- Build branches around collected data, checks, rules, and review steps
- Run a hosted flow, an embedded SDK, or a backend onboarding request
Percy investigation agentsCase research and review assistance+
- Investigate watchlist hits, adverse media, businesses, and transaction alerts
- Record supporting sources and task history; test agent changes against resolved cases
Review PlatformHuman review operations+
- Route review types into queues with assigned teams and service targets
- Prioritize by due date, age, and base priority; track assigned and overdue work
Vault and Vault ProxySensitive-data handling+
- Store structured identity fields and custom data with field-level decryption access
- Search fingerprinted fields and pass vaulted data to configured destinations
Bank Account LinkingConnected account evidence+
- Link accounts through data aggregators and vault the returned account information
- Retrieve balances and transactions; use ownership information in identity-risk checks
Onboarding risk signalsDecision evidence+
- Retrieve reason codes, notes, severity, and timestamps for a specific onboarding
- Expose signal scopes for identity, device, behavior, documents, and business attributes
Trust FabricInvestigation context and agent governance+
- Connect related customer, case, and agent signals for investigation
- Retain provenance for shared precedents and require human approval of proposed memory
SentiLink fraud checkConnected synthetic-identity detection+
- Add the documented SentiLink check to onboarding
- Use its resulting risk signals in downstream policy rules
AI capabilities
Percy agents with versioned workflow actions
Percy supports evidence gathering and case analysis. A playbook can run a released agent version and pass its structured output to rules and action nodes. Those nodes can set onboarding decisions, request review, or update status, tags, and notes. This documents bounded workflow automation, not unrestricted authority to clear compliance cases, move funds, or file reports.
What to validate
Inspect the pinned agent version, allowed inputs, downstream actions, and approval gates. Test conflicting evidence and mandatory review paths. Organizational-memory approval is documented; confirm the separate permissions for case disposition and policy changes. Exact matching and workflow rules do not establish predictive fraud-model performance.
Implementation
Integration checklist
- Map user, business, onboarding, and external identifiers separately. Keep secret keys on the backend and distinguish sandbox from production credentials.
- Handle pass, fail, pending, incomplete, none, and error. Evaluate the manual-review flag separately; a frontend completion callback is not sufficient authorization to admit a customer.
- Validate SDK completion tokens on the backend. Verify webhook signatures, process final decisions, and re-fetch status after manual review.
- Confirm account-specific idempotency behavior before retries. The guide describes onboarding_external_id while the preview reference omits it and gives conflicting external_id guidance.
- Test field access, proxy destinations, data export, deletion, and recovery. Separate retained vault data from temporary onboarding data and versioned agent output.
Commercial scope
Quote required. Confirm charges for onboarding checks, connected providers, bank linking, vault storage and access, review users, Percy executions, support, and implementation. No public unit rate or included module entitlement is assumed.
Questions for the demo
- Which KYC, KYB, watchlist, bank-data, and fraud providers are enabled, and which need separate licenses?
- Which Percy actions run automatically, which require approval, and can we replay the evidence and exact policy version?
- Which preview APIs are enabled for our account, and what are the current idempotency, webhook retry, and event-ordering contracts?
- Can we export vault data, review decisions, agent traces, and organizational memory in a usable format?
- Which controls must remain in our transaction-monitoring, payment, underwriting, and regulatory-reporting systems?
Engineering
- Inputs
- Existing person or business identifier, playbook key, and declared prerequisite data · Structured identity fields and custom sensitive-data fields · User-authorized bank account connection · Optional initial user vault data
- Outputs
- Onboarding identifier, status, manual-review flag, and optional continuation token/link · Structured agent output made available to configured rules and actions · Authorized decrypted fields and search results for fingerprinted attributes · Connected account, balance, ownership, and transaction information · Footprint user identifier and optional external or sandbox identifiers
- Webhooks
- HTTPS JSON POST with a 2xx acknowledgment. Documented events include footprint.onboarding.completed, footprint.user.manual_review, and footprint.watchlist_check.completed. Follow the linked Svix signature verification guidance.
- Decision timing
- Optional synchronous_timeout_secs is greater than zero and at most 30 seconds. Timeout does not cancel the run; completion continues in the background and is reported by webhook.
- Deployment
- Hosted API
- Dependencies
- Backend secret key via X-Footprint-Secret-Key or HTTP Basic authentication · Separate sandbox and live API keys · Account access to the Preview endpoint and configured playbook · Backend validation of SDK completion tokens · Interactive onboarding session for KYB with beneficial-owner verification · Released agent version pinned in each agent node · Field-scoped decryption permissions
Verified details 25
- Deployment
Hosted API
Source Checked 2026-09-18 - Dependency
Backend secret key via X-Footprint-Secret-Key or HTTP Basic authentication
Source Checked 2026-09-18 - Dependency
Separate sandbox and live API keys
Source Checked 2026-09-18 - Inputs
Existing person or business identifier, playbook key, and declared prerequisite data
Source Checked 2026-09-18 - Outputs
Onboarding identifier, status, manual-review flag, and optional continuation token/link
Source Checked 2026-09-18 - Decision timing
Optional synchronous_timeout_secs is greater than zero and at most 30 seconds. Timeout does not cancel the run; completion continues in the background and is reported by webhook.
Source Checked 2026-09-18 - Dependency
Account access to the Preview endpoint and configured playbook
Source Checked 2026-09-18 - Webhooks
HTTPS JSON POST with a 2xx acknowledgment. Documented events include footprint.onboarding.completed, footprint.user.manual_review, and footprint.watchlist_check.completed. Follow the linked Svix signature verification guidance.
Source Checked 2026-09-18 - Dependency
Backend validation of SDK completion tokens
Source Checked 2026-09-18 - Dependency
Interactive onboarding session for KYB with beneficial-owner verification
Source Checked 2026-09-18 - Outputs
Structured agent output made available to configured rules and actions
Source Checked 2026-09-18 - Connected provider
Middesk
Source Checked 2026-09-18 - Connected provider
Baselayer
Source Checked 2026-09-18 - Connected provider
SentiLink
Source Checked 2026-09-18 - Dependency
Released agent version pinned in each agent node
Source Checked 2026-09-18 - Connected provider
LexisNexis
Source Checked 2026-09-18 - Connected provider
Experian
Source Checked 2026-09-18 - Connected provider
ComplyAdvantage
Source Checked 2026-09-18 - Inputs
Structured identity fields and custom sensitive-data fields
Source Checked 2026-09-18 - Outputs
Authorized decrypted fields and search results for fingerprinted attributes
Source Checked 2026-09-18 - Dependency
Field-scoped decryption permissions
Source Checked 2026-09-18 - Inputs
User-authorized bank account connection
Source Checked 2026-09-18 - Outputs
Connected account, balance, ownership, and transaction information
Source Checked 2026-09-18 - Inputs
Optional initial user vault data
Source Checked 2026-09-18 - Outputs
Footprint user identifier and optional external or sandbox identifiers
Source Checked 2026-09-18
Vendor example
Published vendor example. Confirm the current version and required credentials.
{
"id.first_name": "Jane",
"id.last_name": "Doe",
"custom.user_id": "7c50e2bc-c31f-42e3-b2b0-9852010cfd58"
}Integration limits
- Published vendor payload example, linked to its source. It is not a complete integration or a tested production request.
- Empty country, deployment, or provider lists mean not verified in this review. A documented country refers to the specific product noted in its source, not universal platform coverage.
- Public API behavior can vary by account and version. Sandbox keys do not call production verification vendors. No hosting country or self-hosted deployment was verified.
- The API page is explicitly Preview. The public request example supplies both external_id and fp_id although the field rules prohibit that combination; do not copy that example unchanged.
- Pending verification can finish after the SDK callback. Re-fetch current status after a manual-review event. Retry count, acknowledgment timeout, ordering, and a delivery SLA were not verified from this guide.
- The guide describes onboarding_external_id for onboarding idempotency. The Preview endpoint reference omits it and instead describes external_id as an idempotency ID. Confirm the exact account schema and retry behavior before implementing.
- The API also distinguishes pending from incomplete; incomplete supplies a secret token to resume user input.
- Middesk and Baselayer are documented KYB choices; SentiLink is the synthetic-fraud check. No product-by-country list or included provider contract was verified.
- Named integrations are documented for investigation context. The page does not establish a single license, uniform API coverage, or unrestricted data access.
- The documented JSON write limit is 32KB; larger custom objects use an upload route, with a stated 25MB limit. Confirm storage, retention, export, and recovery requirements.
- Open-banking aggregators supply data. No named aggregator or complete country-availability list is established here. The web flow requires footprint-js 5.0.0 or later; mobile bank linking must run sequentially with onboarding because they share an underlying object.
- The linked vendor example is a JSON request body, not a complete HTTP call. When x-external-id is supplied, this endpoint says initial data cannot also be sent in the body.
Sources 13
Source review dates are shown above. Product claims come from public sources. Fit, limits, and evaluation questions are our analysis. This is not a hands-on performance test. Methodology · Changelog