← Blog

DORA Third-Party Risk for AI: What ICT Third-Party Providers Have to Show

Parminder Singh
Parminder Singh··7 min read
Summarize with AI

DORA took effect January 17, 2025. The regulation treats AI vendors as ICT third-party service providers. Financial entities must maintain a register of contractual arrangements, monitor concentration risk, and demonstrate exit strategies. AI inference sits squarely inside the obligation.

Compliance & Regulationdorathird-party-riskfinancial-servicescomplianceai-governanceict-risk
DORA Third-Party Risk for AI: What ICT Third-Party Providers Have to Show

The EU Digital Operational Resilience Act (Regulation 2022/2554) took effect on January 17, 2025. DORA covers the financial entities listed in Article 2. The regulation treats AI inference vendors as ICT third-party service providers under Article 3(19). Financial entities have to register the relationship, monitor concentration risk, prove an exit strategy works, and submit incident reports through national competent authorities to the European Supervisory Authorities.

I want to walk through what DORA requires from financial entities using AI vendors, how the AI inference relationship fits the third-party framework, and what evidence the joint ESAs expect during a thematic review.

TL;DR

  • DORA treats an AI inference relationship as an ICT third-party arrangement when it meets the Regulation's service definition.
  • The register needs an identifiable service, provider, function, contract, data scope, and a current criticality assessment.
  • A defensible exit plan has a tested alternate route, rather than a termination clause alone.
  • Request records can support a review of HTTP AI traffic that actually traverses the enforcement point.

How DORA classifies AI vendors

DORA defines an ICT service in Article 3(21) as "digital and data services provided through ICT systems to one or more internal or external users." AI inference fits the definition. The financial entity submits a prompt over an HTTPS endpoint and receives a response generated by the vendor's model. The service is digital, runs on ICT systems, and serves users inside the financial entity.

Article 3(19) defines an ICT third-party service provider as an undertaking providing ICT services. Major AI inference vendors and smaller specialist providers can fit that definition. The financial entity that buys their inference services has the third-party provider obligations under Articles 28 to 30 of DORA.

The classification follows the service received, rather than the vendor's primary product line. Microsoft and AWS provide cloud services and AI inference. Each operationally distinct AI service can require its own contractual-arrangement record.

What goes into the register of information

DORA Article 28(3) requires the financial entity to maintain and update a register of information for all contractual arrangements on the use of ICT services provided by third-party providers. The European Banking Authority's operational-resilience materials document the supervisory technical work around those records.

The register covers identification of the third-party provider, the function supported, the criticality classification, the data category processed, the location of data storage and processing, the start date of the contract, the contract reference, and the substitutability assessment. For AI vendors, the register entries get specific. The function supported is "AI inference for customer service summarization" or "AI inference for transaction risk scoring," not "AI." The data category is mapped to the financial entity's data classification scheme, with personal data and confidential data flagged separately.

Financial entities submit the register annually to their national competent authority. The competent authority forwards the data to the European Supervisory Authorities for the EU-wide oversight regime. The ESAs use the register to identify critical ICT third-party service providers that get designated under Article 31 of DORA.

Concentration risk and critical providers

DORA Article 29 requires financial entities to assess concentration risk from ICT third-party arrangements. A high concentration on one vendor for a critical function triggers heightened oversight. The financial entity has to document the substitutability assessment under Article 30(2) and show that the function can move to another provider without unacceptable disruption.

AI inference is a structurally concentrated market. Foundation model capability sits with a small number of vendors. Substitutability between OpenAI and Anthropic for a specific workload requires the financial entity to run model comparison testing, validate output quality, retrain prompt templates, and adjust downstream processing. The substitutability assessment is operationally meaningful, not a paperwork exercise.

The ESAs designate the most material ICT third-party service providers as critical under Article 31. Designation brings the provider under direct ESA oversight. That oversight includes inspections and information requests. The European Banking Authority's operational-resilience materials track the implementation work. Cloud providers and AI vendors are prominent candidate categories.

Exit strategies and the testing requirement

DORA Article 28(8) requires financial entities to maintain exit strategies for ICT third-party arrangements supporting critical or important functions. The exit strategy must be documented, kept current, and tested.

For AI vendors, the exit strategy answers what happens if the vendor terminates the service, suffers a major outage, or fails the criticality test. The strategy covers the alternative provider, the cutover procedure, the data migration path, and the operational continuity arrangement. The financial entity must demonstrate that the strategy works in practice. The Article 24 testing program covers ICT third-party exit scenarios alongside other resilience tests.

DORA Article 28(8) calls for comprehensive, documented exit plans for ICT services that support critical or important functions; those plans must be sufficiently tested and reviewed periodically. For an AI inference service, a contract clause on migration assistance records the vendor's promise. A dated cutover exercise can show whether the financial entity can switch providers and restore the workload within its planned recovery time. Keep the contract clause and the exercise result as separate evidence.

Third-party risk evidence for an AI service

Article 28 makes the financial entity responsible for managing ICT third-party risk throughout the relationship. A useful AI-service review joins the contract record to operating evidence instead of leaving the register as a procurement snapshot.

  • Service record: name the model endpoint, provider legal entity, supported business function, data category, hosting and processing locations, and the contractual reference.
  • Criticality assessment: document the function's disruption impact, dependencies, substitutability result, and the concentration exposure created by the provider.
  • Contract review: record the Article 30 terms that matter to the service, including audit and access rights, incident support, service levels, termination assistance, and data-location commitments.
  • Exit exercise: retain the scenario, alternate route, test date, measured recovery result, defects found, and accountable owner for remediation.
  • Request evidence: preserve the authenticated caller, selected provider route, data classification, policy decision, outcome, and timestamp for HTTP AI requests that traverse the enforcement point.

The fifth record is narrower than the full DORA control set. It gives a review team evidence about routed AI traffic. Application owners still hold evidence for local processing, tool access, and actions after the model response.

Incident reporting and what AI incidents trigger

DORA Article 17 requires financial entities to report major ICT-related incidents to the competent authority. The classification of an incident as major follows criteria laid out in the Joint ESA Regulatory Technical Standard on classification. Geographic spread, duration, number of clients affected, data losses, reputational impact, and economic impact all factor in.

For AI vendors, incidents that trigger reporting include unavailability of the AI service during a critical operational window, unauthorized disclosure of customer data submitted through the AI service, model output that produces material financial harm to customers, and corruption of the AI vendor's underlying infrastructure that affects the financial entity's operations.

Article 19 sets staged reporting for major ICT-related incidents, with detailed timing and content established by the applicable technical standards and national competent authority process. Financial entities need enough service and request evidence to assess exposure quickly after an AI-vendor incident. Prompt-level visibility for routed traffic supports that assessment; it does not replace the entity's broader incident-response process.

The audit trail DORA actually expects

DORA Article 5 assigns the management body ultimate responsibility for the digital operational resilience strategy. The ICT risk management framework is set out in Article 6. Article 9 covers identification, classification, and documentation of ICT-supported business functions. Article 13 addresses monitoring and control of ICT systems.

For AI usage, the operational records expected during a competent authority inspection include the AI vendor used per request, the workforce member or automated agent that initiated the request, the data category exposed, the policy that governed the decision, the outcome of the decision, and the timestamp. The records support AI audit-trail design, the register entries, concentration-risk assessment, exit-strategy testing, and incident reporting.

Application logs often omit important dimensions of the AI-service record. An enforcement point at the AI request boundary can produce evidence for routed HTTP requests, while the application remains the system of record for its own workflow and downstream actions.

DeepInspect

DeepInspect provides an enforcement point for HTTP AI traffic between authenticated users or agents and LLM endpoints. Per-route policies can apply the service classification to a routed request before it reaches the selected model. Each decision produces an audit record containing identity, role, data category, vendor route, policy version, outcome, and timestamp.

Those records can support the AI-specific evidence portion of a DORA third-party risk review. Vendor-route data contributes to concentration analysis. The records also help an entity investigate an incident affecting routed AI calls. They are not a substitute for the complete register, contractual evidence, resilience tests, or application records DORA requires.

If an AI-service register depends only on a quarterly survey of application teams, it will lag the actual provider routes in use. Book a demo today.

Frequently asked questions

Does DORA apply to AI vendors directly or only to financial entities?

DORA applies to the financial entity. The financial entity has obligations regarding the AI vendor relationship under Articles 28 to 30. If the European Supervisory Authorities designate the AI vendor as a critical ICT third-party service provider under Article 31, the vendor itself becomes subject to direct ESA oversight. The first round of designations is expected in 2026. Cloud providers and the largest AI vendors are the categories most likely to see designations.

How does DORA interact with the EU AI Act for financial entities?

The two regulations apply concurrently. DORA covers operational resilience and third-party risk. The EU AI Act covers AI-specific obligations for high-risk systems. A credit scoring system operated by a bank using a third-party AI service falls under both regimes. The DORA register entry covers the third-party relationship. The EU AI Act high-risk classification under Annex III covers the algorithmic obligations including Article 12 record-keeping. Both regimes require operational records at the AI request boundary, which can be produced from a single per-request audit trail.

What is the 4-hour notification window for ICT incidents?

DORA Article 19(4) requires financial entities to submit an initial notification of a major ICT incident to the competent authority within 4 hours of classification as major. Classification can take up to 24 hours under the implementing standard. The 4-hour clock starts from classification, not from incident onset. Financial entities should design their incident response to support the classification decision within hours of detection and to produce the initial notification immediately after classification. AI integrations that lack prompt-level visibility extend the detection window and compress the response window.

Are AI inference services automatically critical functions under DORA?

DORA Article 3(22) defines critical or important functions as functions whose disruption would materially impair the financial entity's compliance with regulatory obligations, financial performance, or soundness or continuity of services. Whether a specific AI inference workload is critical depends on the function the workload supports. AI inference for customer service summarization is typically not critical. AI inference for transaction risk scoring on payment authorization may be critical. The classification is the financial entity's responsibility under Article 8, with the criteria documented and the assessment kept current.

What records does a DORA inspection actually request?

A competent authority inspection can request the register of information, ICT risk-management documentation, the business-impact analysis for the reviewed function, the provider contract, substitutability assessment, exit strategy with recent test results, and operational logs for the relevant period. For AI usage, those logs can include prompt-level records that show who used the AI service, for what purpose, and under which policy. Signed AI audit logs make the integrity requirement concrete. A review can run for several weeks, and inability to produce the records can become its own finding.