BBVA AI Strategy: Building an AI-Native Bank, Agent by Agent

BBVA aims to become an AI-native bank, and it is the top-ranked eurozone bank in the Evident AI Index, in 13th place. It has organized its strategy around eight agentic priorities and published a dated rollout plan through 2027. It reports many process results, but no group-level financial value from AI.

The ambition is an AI-native bank. Announcing its OpenAI alliance in December 2025, BBVA said it is building an AI-native banking model with a radically customer-centric proposition. It is one of the clearest bank examples of the shift toward AI native organizations.

BBVA names both the opportunity and the threat. Its October 2026 AI strategy presentation says AI removes bottlenecks and gives first movers an edge.

It also warns that new competitors could use AI agents to come between the bank and its customers.

The focus is “The Eight,” which BBVA calls its first bank-wide agentic agenda. It covers:

• clients: a personal advisor called Blue, and an AI Banker
• operations: AI risk management, process redesign and software development
• employees: AI for every employee, through an “AI Alter Ego”
• foundations: a data-rich company, and AI architecture and technology

The primary focus is the customer. Among the banks in this index so far, BBVA is the only one whose main AI goal is customer experience rather than efficiency.

BBVA has not said publicly what it will not prioritize.

The plan runs in three steps. BBVA will prove value through The Eight, then industrialize through “The Frame,” a common model for building, deploying, managing and governing agents. The third step is scaling impact by redesigning processes end to end.

Agents are already in production. BBVA reports AI agents in claims, recoveries and financial crime, and contact-centre agents in its core markets. The agentic version of Blue is live in Peru, with Mexico planned for the fourth quarter of 2026. Agents work with humans in the loop.

The main technology partner is OpenAI. Under a multi-year alliance, every one of BBVA’s more than 120,000 employees has ChatGPT Enterprise. BBVA also offers Google’s Gemini to staff, and it stresses keeping its options open across AI providers. About 4,000 data specialists work on a single global data platform.

The workforce stance is about shifting work. BBVA says employees will move toward strategizing and supervising rather than executing. It puts the scale of work affected at about 10,000 roles in banking operations and 12,000 developers.

It has not disclosed headcount cuts tied to AI.

Enablement starts at the top. BBVA trained 250 leaders, including its CEO and chair, and has a Head of Global AI Adoption, according to an OpenAI case study. It also set up an internal AI network to move proofs of concept into production faster, as covered in OpenAI’s 23 things every AI leader must do.

Staff must train before they get tools. Its AI Express training is mandatory before staff get access to Gemini. More than 105,000 employees took it in its first week, and AI training passed 280,000 hours in 2025.

Ownership sits at the top of the organization. In June 2026, BBVA created a global AI Transformation area led by Antonio Bravo, reporting at the top level of the bank. Chair Carlos Torres Vila is the public sponsor of the OpenAI alliance. No AI-specific investment figure has been disclosed.

Governance is built into the agent model. BBVA publishes eight responsible AI principles, covering transparency, security, privacy and human oversight.

The Frame builds in guardrails, evaluation, monitoring and value management by design. In BBVA’s words, control is mandatory to scale AI.

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

• Fluency floor: Level 1 (Chatting). Training is mandatory before tool access, which sets a usage baseline but not a work-quality standard.
• Highest level evidenced: Level 3 (Integrating), company-claimed. BBVA reports about 80% less time to register customer claims and about 30% less time to produce enterprise financial programs. Voice agents resolve about 90% of interactions, but only in pilots, so Level 4 is close but not yet evidenced at scale.
• Portfolio design: Multi-level, with leader training, dedicated adoption roles and role-specific agents. Some role targets are disclosed: AI for every risk analyst in corporate and investment banking in 2026, and for commercial banking in 2027.

BBVA ties each agent to its value. Its “value management” approach links every agent to its costs and expected value. It tracks time to serve, resolution rates, process cycle times, developer productivity and adoption.

It publishes process results. The October 2026 presentation reports the following, all company-claimed:

• about 90% successful resolution in pilot voice interactions
• Blue handling more than 50% of in-app chats between clients and relationship managers
• about 80% less time to register customer claims
• 28% higher developer productivity

Adoption is broad. BBVA says 75% of employees use AI tools regularly, saving 2.4 hours a week each. Both figures are self-reported.

The roadmap is dated through 2027. Contact-centre agents reach Mexico in the fourth quarter of 2026. In 2027, agentic Blue reaches retail and commercial customers in Spain and Mexico, and employee “Alter Egos” begin redesigning end-to-end processes.

One figure is missing: a group-level financial value for AI.

Change log

• July 2025: Mandatory AI training is linked to Gemini access.
• December 2025: The multi-year OpenAI alliance brings ChatGPT Enterprise to all employees.
• June 2026: A top-level AI Transformation area is created.
• October 2026: The AI strategy presentation sets out a 2027 roadmap, and Evident ranks BBVA 13th, first in the eurozone.

• Disclosure: 4 of 4. BBVA publicly answers all four questions, with a named agenda, a dated roadmap and measured process results.
• It is the clearest customer-first AI strategy in banking. Blue, the AI Banker and the warning about agent intermediaries all point outward, not only at cost.
• It is the furthest along on agents. Agents run in claims, recoveries, financial crime and contact centres, with humans in the loop.
• The say-do gap is in money and scale. Several headline results come from pilots or self-reported surveys, and no group-level financial value has been published.
• Watch 2027. Agentic Blue in Spain and Mexico will be the real test of the customer strategy.

What the numbers show: Revenue per employee rose 36%, but part of that is inflation in Turkey and Argentina, and headcount still grew 10%. The numbers do not yet show a workforce effect from AI.

Revenue, headcount and revenue per employee, FY2022 to FY2025

• FY2022: €24.74bn revenue, 115,675 employees, €214k per employee
• FY2023: €29.54bn revenue, 121,486 employees, €243k per employee
• FY2024: €35.48bn revenue, 125,916 employees, €282k per employee
• FY2025: €36.93bn revenue, 127,174 employees, €290k per employee

Change, FY2022 to FY2025: revenue up 49%, headcount up 10%, revenue per employee up 36%.

Notes: Revenue is BBVA’s gross income, which it reports as total revenues. Headcount is at year-end. 2022 is restated for IFRS 17. Part of the revenue growth reflects high inflation in Turkey and Argentina. Trend figures stay in € so exchange-rate moves do not distort them. At the 2025 average exchange rate, latest revenue per employee is about $328k.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: BBVA quarterly results reports and 20-F filings.

What is BBVA’s AI strategy?

BBVA aims to become an AI-native bank. Its strategy is built around “The Eight,” a bank-wide agentic agenda covering clients, operations, employees, data and technology. It is delivered through a common model for building and governing agents.

‍

Does BBVA use OpenAI?

Yes. Under a multi-year alliance announced in December 2025, all of BBVA’s more than 120,000 employees have ChatGPT Enterprise, and the two companies work together on customer and banker tools. BBVA also offers Google’s Gemini to staff.

‍

What results has BBVA reported from AI?

BBVA reports about 80% less time to register customer claims, 28% higher developer productivity and about 90% resolution in pilot voice interactions. These are company figures, and some come from pilots.

‍

Is BBVA cutting jobs because of AI?

BBVA has not announced AI-related job cuts. It says employees will shift from executing work to strategizing and supervising it.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings materials, filings and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix. Company figures are marked as company-claimed unless they have been independently verified.