Join our Newsletter — 33% off our NHI Course
Home› Guides› Relying Party Integration with Digital Identity Wallets Guide:…
Guide Customer Identity

Relying Party Integration with Digital Identity Wallets Guide: Registration, Requests, Verification and Fraud Controls

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 27 October 2026 14 min read
On this page

Accepting a credential from a digital identity wallet is not the same as adding another login button: the relying party takes on registration duties, certificate management, cryptographic verification and new fraud risks. Banks, telecoms operators, marketplaces and SaaS providers are being asked to accept the EU Digital Identity Wallet, mobile driving licences and other wallet credentials, often because regulation or customers expect it. This guide is for the engineering, identity and risk teams on the relying party side. It covers registration and certificates in the EUDI model, how to build a minimal request, the verification steps a relying party must perform, the threats and fraud patterns that remain, operations and logging, and a phased rollout. For the wallet landscape itself, see the Digital Identity Wallets Guide.

Key takeaways

  • Registration comes before integration: in the EUDI ecosystem a relying party registers with a Registrar in its Member State, receives access certificates for each relying party instance and registration certificates for each intended use.
  • Ask only for what you registered and need: wallets compare your request with your registration certificate and warn the user about any extra attribute, so over-asking damages conversion and invites complaints to data protection authorities.
  • Verification is your job: issuer signature against a trusted list, revocation status, device binding (key binding or mdoc authentication) and nonce and audience checks must all pass before you act on attributes.
  • Your relying party keys are crown jewels: the access certificate private key lets anyone request citizens' data in your name, so it belongs in an HSM with tight access, rotation and monitoring.
  • Wallets reduce but do not end fraud: cross-device phishing, social engineering of the holder, and weak fallback channels remain; design session binding and fallbacks deliberately.

What the relying party must put in place

The Architecture and Reference Framework (ARF) for the European Digital Identity Wallet, version 3.0.0 of 21 July 2026 at the time of writing, sets out the relying party lifecycle in its trust model chapter. Other wallet ecosystems, such as mobile driving licence programmes based on ISO/IEC 18013-5, differ in detail but need the same building blocks. See the EUDI Wallet Architecture and Reference Framework Guide.

Building blockWhat it isRelying party action
Registration with a RegistrarRegistration in the Member State where the relying party is established; the Registrar assigns an EU-wide unique relying party identifier and publishes the entry in a public registryRegister each service and intended use; keep contact and privacy details current
Intended usesEach service registers one or more intended uses with the attributes it will request; identifiers are generated by the RegistrarMap every customer journey to a registered intended use
Registration certificateIssued per intended use by a Provider of registration certificates, listing the registered attributes; revocation is published as a status list under ETSI TS 119 475Include the right certificate with each presentation request
Access certificateIssued by an Access Certificate Authority to each relying party instance; used to sign presentation requests so wallets can authenticate youProtect the private key, manage expiry and revocation
Relying party instanceThe hardware and software that talks to wallets; a relying party can run several (production, fallback, proximity readers)One access certificate per instance; inventory them
Trust anchorsTrusted Lists and Lists of Trusted Entities (LoTE) for PID providers, attestation providers and access CAsDownload regularly and distribute to every instance

Two points are easy to miss. First, the ARF states that registration is for transparency: the Registrar does not pre-authorise you to request the attributes you list, and if a legal basis or authorisation is needed it sits outside the wallet ecosystem. Second, a Registrar can suspend or cancel a registration, which revokes all access and registration certificates and stops your service working with wallets, so treat registration hygiene as an availability risk.

Intermediaries: if a vendor requests wallet data on your behalf, the vendor registers as a relying party in its own right, and your registration certificate names it in a usesIntermediary field. The wallet then shows your trade name, not the intermediary's. The ARF requires intermediaries to delete data obtained from a wallet immediately after passing it to you, and data deletion requests always go to you, not the intermediary. Contract for that, and audit it. See the Third-Party Risk Management Guide.

Building a minimal presentation request

For remote flows the main protocol is OpenID for Verifiable Presentations (OpenID4VP 1.0, Final, 9 July 2025); for proximity flows it is ISO/IEC 18013-5. The OpenID4VC High Assurance Interoperability Profile (HAIP 1.0, Final, 24 December 2025) narrows OpenID4VP for high assurance use. See the OpenID for Verifiable Presentations (OpenID4VP) Guide.

  • Use DCQL and request claims, not credentials: a dcql_query names the credential type and the specific claims you need. Use credential_sets and claim_sets to accept alternatives rather than asking for everything.
  • Prefer derived attributes: for age checks request an over-18 style attribute where the credential offers one, not the date of birth. See the Age Assurance Guide.
  • One intended use per request: the ARF expects a single registration certificate per request; if a journey has two purposes, send two requests.
  • Sign requests: send a signed request object (typ oauth-authz-req+jwt) by request_uri. HAIP requires the x509_hash client identifier prefix for signed requests, and the signing certificate must not be self-signed.
  • Fresh nonce every time: the nonce is required, must be cryptographically random and is what binds the presentation to your request.
  • Encrypt the response: use direct_post.jwt (or dc_api.jwt with the browser API), with ephemeral ECDH-ES P-256 keys per request, and support A128GCM and A256GCM as HAIP requires.
  • Require holder binding: require_cryptographic_holder_binding defaults to true in DCQL; keep it, so a copied credential cannot be replayed.
  • Use transaction data for payments and signing: transaction_data binds the user's approval to a specific transaction; wallets reject types they do not know, so agree the type in the relevant rulebook.

Request and format reference

ItemSD-JWT VCISO mdoc
Format identifier in DCQLdc+sd-jwtmso_mdoc
Issuer signature key resolution (HAIP)X.509 via x5c headerIssuer certificate in the mobile security object
Device binding proofKey binding JWT (KB-JWT), always present when holder binding appliesmdoc authentication (deviceSignature or MAC)
RevocationToken Status List (status_list)Optional; if used, a mechanism from ISO/IEC 18013-5 second edition
Minimum algorithm (HAIP)ES256ES256

See the SD-JWT and SD-JWT VC Guide and the Mobile Driving Licence (mDL) Guide for format detail.

Verifying what the wallet returns

The ARF lists the checks a relying party instance performs on every presentation. Several are described as recommended rather than mandatory at ecosystem level; for any regulated or high-value use, treat all of them as mandatory in your own policy.

  1. Decrypt and match the session: decrypt the response with the per-request key, confirm it answers your request (nonce, audience, and for redirect flows that it arrives in the same user session).
  2. Verify the issuer signature: validate against a trust anchor for the right category: PIDs, qualified attestations (QEAAs), public sector attestations (PuB-EAAs) or non-qualified EAAs. Store trust anchors so you can tell the categories apart; a valid signature from the wrong category is a failure.
  3. Check the issuer is entitled: optionally confirm the attestation provider is registered to issue that attestation type.
  4. Check revocation: credentials valid for more than 24 hours carry a status list or revocation list reference. If no fresh status is available, the ARF expects a documented risk decision; for high-value actions, refuse. See the Token Status List and Credential Revocation Guide.
  5. Verify device binding: check the KB-JWT or mdoc authentication signature over your nonce. The ARF notes that trusting the wallet's own user authentication as user binding requires the relying party to verify device binding.
  6. Decide on extra user binding: for unsupervised remote journeys with high risk, comparing the portrait to a live selfie needs presentation attack detection and injection attack detection, including deepfake detection. See the Deepfake-Resistant Identity Verification Guide.
  7. Check combined presentations: when attributes come from two or more attestations, confirm they belong to the same user, using cryptographic binding where the ecosystem supports it, not just matching names.
  8. Apply business rules: validity dates, attribute values, and your own fraud signals, then record the outcome.

Threats and fraud patterns that remain

ThreatHow it worksRelying party control
Cross-device phishingA fake site relays a genuine relying party QR code so the victim presents a credential that completes the attacker's sessionPrefer same-device flows; HAIP requires rejecting responses where the redirect does not return to the same session; use the browser Digital Credentials API where available. See the Cross-Device Flow Security Guide
ReplayA captured presentation is resentFresh nonce per request, single use, short validity, holder binding
Relying party impersonationAttacker poses as your service to harvest attributesAccess certificate key protection; wallets reject requests whose chain fails; monitor for lookalike services
Coerced or socially engineered holderVictim is talked into presenting a credential for a "refund" or "account check"Clear purpose text, transaction data for payments, velocity and behavioural checks
Stolen or compromised deviceUnlocked phone used to present credentialsRely on wallet user authentication plus your own risk signals for high-value changes
Weak fallbackAttackers skip the wallet and use the document upload or call centre pathFallback at equivalent assurance; monitor fallback share. See the Account Recovery and Help Desk Security Guide
Over-collectionRequests beyond the registration certificate trigger warnings and complaintsRequest review against registration; automated tests on DCQL

For a deeper treatment of wallet fraud, see the Digital Identity Wallet Fraud Risks Guide.

Where identity fits

Every relying party instance holds an access certificate and private key, a non-human identity that can request citizens' data. Keep keys in an HSM or managed key service, never in application configuration; one key per instance; named owners; expiry alerts well before renewal; and immediate revocation through the Access CA if an instance is compromised. Response encryption keys must be ephemeral per request, not a static key copied across servers. See the Machine Identity, PKI and Certificate Lifecycle Guide.

Privacy, user rights and data handling

  • The wallet keeps a log of your requests: the ARF requires a transaction log listing the attribute identifiers requested and presented, and lets users report suspicious requests to a data protection authority even if nothing was presented.
  • Deletion requests come through your existing channels: wallets open the web form, email or phone contact in your registration certificate. Make sure those channels are monitored and linked to your GDPR process.
  • Store the minimum: keep the verification outcome, credential type, issuer, status check result and a reference, not the full presentation, unless law requires retention.
  • Very large online platforms: the ARF notes that Article 5f(3) of the European Digital Identity Regulation requires them to accept the wallet for user authentication only on the user's voluntary request and for the minimum data necessary.
  • Strong user authentication: Article 5f(2) uses the term strong user authentication; the ARF assumes wallets meet PSD2 strong customer authentication requirements, which matters for banks considering wallets for login and payment approval.

Decision guidance

  • Build or use an intermediary: an intermediary removes certificate and protocol work, which suits small relying parties; you still remain the relying party in law, and you add a processor whose deletion duties you must verify. Large banks and platforms usually build, to control keys, logging and fraud signals.
  • Redirect or browser API: on the web, the W3C Digital Credentials API gives the browser and operating system a role in binding the request to the origin, which helps against cross-device phishing; the redirect flow with direct_post.jwt covers wallets and browsers without it. See the W3C Digital Credentials API Guide.
  • Which credentials to accept: PIDs for regulated onboarding, mDL or other attestations for age and licence checks, and qualified attestations where you need an attribute with legal effect. Configure trust anchors per category, not one global list.
  • Wallet first or wallet optional: keep an equivalent alternative while wallet adoption grows; set a target share rather than forcing the wallet.

Implementation roadmap

  1. Scope intended uses: list journeys, the attributes each needs, the legal basis and retention; review with privacy and legal.
  2. Register: register services and intended uses with your Registrar, obtain access certificates per instance and registration certificates per intended use, and record them in your certificate inventory.
  3. Build the verifier: implement OpenID4VP with DCQL, signed requests, encrypted responses and the verification steps above; add ISO/IEC 18013-5 for proximity if needed. Test against reference wallets and conformance suites where available.
  4. Trust anchor automation: schedule downloads of Trusted Lists and LoTEs, validate them, and push updates to every instance; alert on failures.
  5. Pilot: offer the wallet option to a small share of customers alongside the existing journey; compare completion, fraud and support contacts.
  6. Roll out and keep a rollback: expand by segment; if verification error rates spike or a trust list update breaks validation, route users back to the existing journey with a feature flag while you fix it.
  7. Operate: renew certificates ahead of expiry, review requests against registrations after every product change, and handle data protection complaints.

Detection and measurement

Log per presentation: relying party instance, intended use, credential type and issuer, verification step outcomes (signature, status, device binding, session match), nonce reuse attempts and the final decision. Do not log attribute values.

  • Verification failure rate by reason: signature, revocation unavailable, device binding missing, session mismatch.
  • Revocation check coverage: percentage of presentations with a fresh status check (target 100 per cent for high-value actions).
  • Over-request rate: requests containing attributes not in the registration certificate (target zero).
  • Fallback share and fallback fraud rate: to show attackers shifting to the weaker path.
  • Certificate health: days to expiry for every access and registration certificate, and time to update trust anchors after a list change.

Framework mapping

FrameworkReferenceRelevance
NIST SP 800-53 Rev. 5IA-8, IA-8(2), IA-8(4)Accepting external credentials and defined profiles (HAIP, rulebooks)
NIST SP 800-53 Rev. 5IA-5(2), SC-17Certificate validation, trust anchors and relying party certificates
NIST SP 800-53 Rev. 5SC-23, PT-3, SI-12Session binding, purpose limitation, retention of presented data
NIST CSF 2.0PR.AA-02, PR.AA-03Proofed identities bound to credentials; authentication of users and services
ISO/IEC 27001:2022 Annex AA.5.34, A.8.24, A.5.19PII protection, cryptographic key management, supplier relationships for intermediaries

Common failure patterns

  • Skipping revocation and device binding checks because the ecosystem calls them recommended: a revoked or copied credential then passes.
  • One trust list for everything: accepting a non-qualified attestation signature as if it were a PID provider signature.
  • Requesting the full credential: asking for every PID attribute when the journey needs age or name only, which triggers wallet warnings and complaints.
  • Static keys in configuration: access certificate private keys or response encryption keys copied into application settings and containers.
  • Cross-device flows without session binding: accepting a response that arrives in a different browser session from the one that started the request.
  • Forgetting the intermediary: no contract or audit of the intermediary's immediate deletion duty.

Practitioner checklist

  • Register every service and intended use, and map each customer journey to exactly one intended use.
  • Hold access certificate private keys in an HSM, one per relying party instance, with owners and expiry alerts.
  • Build DCQL requests that ask only for registered, necessary claims, preferring derived attributes.
  • Sign requests, use fresh nonces, encrypt responses with ephemeral keys and require holder binding.
  • Verify issuer signature by trust anchor category, revocation status, device binding and session match on every presentation.
  • Automate Trusted List and LoTE downloads to all instances and alert on failures.
  • Keep an equivalent-assurance fallback and monitor its fraud rate.
  • Store outcomes, not attribute payloads, and route wallet deletion requests into your GDPR process.

Standards and references

Related NHI Mgmt Group resources: Digital Identity Wallets Guide · eIDAS 2.0 and the European Digital Identity Framework Guide · Wallet Unit Attestation and Device-Bound Keys Guide · Digital Identity Wallet Fraud Risks Guide

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 27 October 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.