In-house Financial Services Spain Company newsroom

BBVA

BBVA needed to prove that AI agents could initiate consumer purchases securely within existing payment infrastructure while preserving cardholder consent, issuer oversight, and EU Strong Customer Authentication compliance.

Reported impact

Business outcome

BBVA completed its first live AI agent-initiated card transaction and showed that agentic payments can run in a real merchant environment using existing infrastructure, cardholder consent controls, and current regulatory protections.

  • Compliance
  • Customer Experience
  • Scalability

Solution

BBVA used Visa Intelligent Commerce in a live agentic commerce test to let an AI agent initiate a purchase on behalf of a cardholder using real card credentials and an active merchant environment. The transaction was processed on existing payment infrastructure rather than a simulated sandbox, with tokenisation protecting credentials as the payment request entered the flow and real-time fraud monitoring checking the transaction before authorisation. To satisfy European Strong Customer Authentication requirements, BBVA added Visa Payment Passkeys so the cardholder could approve the online payment biometrically instead of by password or SMS. The implementation linked AI systems, merchant systems, issuer controls and Visa security protections into a single production-style path, showing that agent-initiated payments can be served securely and compliantly at scale even though the source does not describe storage internals or model architecture.

Data flow

An AI agent initiated a purchase request for a cardholder, the request was routed through Visa Intelligent Commerce using tokenised card credentials and real-time fraud monitoring, and the cardholder then authenticated the payment with Visa Payment Passkeys to satisfy SCA before the transaction was completed with the active merchant and issuer systems.

Solution architecture

4 components · 2 layers
  1. Orchestration
    • Visa Intelligent Commerce Enabled the live agentic commerce transaction and connected the AI-initiated purchase to the secure payment infrastructure.
  2. Governance
    • tokenisation Protected the card credentials used in the AI-initiated payment flow.
    • real-time fraud monitoring Checked the transaction in real time to keep the agent-initiated payment secure.
    • Visa Payment Passkeys Provided biometric cardholder authentication so the payment could meet SCA requirements without passwords or SMS codes.

Architecture clues

  • Biometric passkeys were used to meet EU Strong Customer Authentication requirements
  • Live agentic commerce activity under Visa's Agentic Ready programme in Europe
  • Source describes issuer oversight, cardholder consent, and existing protections as part of the operating model
  • Used real card credentials rather than a simulated payment setup

Evidence from the source

"BBVA has completed a transaction initiated by an artificial intelligence agent on behalf of a cardholder, as part of a live agentic commerce activity"
"The payment was carried out using real card credentials and the systems of an active merchant"
"The transaction successfully tested by BBVA was enabled through Visa Intelligent Commerce, using the same technologies that support secure digital payments, including tokenisation and real-time fraud monitoring."
"This demonstrates that AI agent-initiated transactions can operate within existing regulatory frameworks"
"fully maintaining cardholder consent, issuer oversight and compliance with existing protections."
"the transaction also used Visa Payment Passkeys, a biometric authentication method that enables consumers to authorise online payments securely without passwords or SMS codes."
Transparency Evidence & methodology 71/100 reference value · 92% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value71/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth25/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity16/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact0/25
    0 metrics describe measurable impact.
  • Deployment scale10/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence92%
  • Source authority25/25
    The deployment is described by the adopting organization or a government source.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence20/20
    6 direct source quotes retained.
  • Date provenance10/10
    The publication date comes from the source page or publisher.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity8/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. BBVA

    BBVA customer story

    First-party
    https://www.bbva.com/en/innovation/bbva-completes-its-first-ai-agent-initiated-transaction-together-with-visa/
    Source type
    Customer-published story
    Published
    Jul 2, 2026
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date

Date evidence

Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.