Most enterprise AI announcements follow a familiar script: a pilot program, a handful of enthusiastic employees, a press release full of vague efficiency claims. BBVA just did something genuinely different. The Spanish multinational bank has rolled out ChatGPT Enterprise to its entire global workforce — all 100,000 employees — and formalized a direct strategic partnership with OpenAI to build AI-powered banking products from the ground up. That’s not a pilot. That’s a full organizational bet, and it’s worth understanding exactly what it involves and what it signals for the rest of the financial industry.
How BBVA Got Here: A Bank That’s Been Serious About Tech for Years
BBVA isn’t a typical bank that suddenly discovered AI in 2023. The Madrid-based institution has been positioning itself as a technology company that happens to do banking for well over a decade. It was an early investor in fintech, built its own digital banking platform from scratch in several markets, and has consistently outspent European peers on tech infrastructure as a share of revenue.
That context matters because it explains why a 100,000-seat ChatGPT Enterprise deployment isn’t as surprising as it sounds coming from BBVA — but it’s still a massive operational undertaking that most financial institutions would balk at. Compliance requirements, data residency rules, customer privacy obligations, and internal security protocols make wide-scale AI deployment in banking genuinely complicated. BBVA apparently worked through all of that.
The bank started experimenting with OpenAI tools in 2023, initially with smaller internal teams. By early 2024 it had expanded access significantly. The full 100,000-seat deployment and formal partnership announcement, detailed by OpenAI in June 2025, represents the culmination of roughly two years of internal scaling and governance work. That timeline is actually fast for a regulated financial institution operating across more than 25 countries.
What the BBVA-OpenAI Partnership Actually Covers
There are two distinct components here, and conflating them misses what’s interesting about this deal.
The first is the ChatGPT Enterprise deployment — giving every BBVA employee access to OpenAI’s business-grade AI assistant. This is the productivity layer: helping analysts draft reports faster, letting customer-facing staff prepare for complex client conversations, enabling developers to write and review code more efficiently, and giving HR teams better tools for internal documentation. Standard enterprise AI use cases, but at a scale most companies haven’t hit.
The second component is the strategic partnership itself, which goes beyond a software license. According to OpenAI’s announcement, the two organizations are collaborating on building AI-native banking products — meaning new features and services that are designed around AI capabilities from the start, not retrofitted onto legacy systems. Think AI-powered financial planning tools, smarter fraud detection interfaces, and client-facing products that use language models to handle genuinely complex financial queries.
Key elements of what’s been deployed and built:
- ChatGPT Enterprise at full scale — all 100,000 staff across BBVA’s global footprint have access, with the security and data privacy controls enterprise deployments include (conversations aren’t used to train OpenAI models, data stays within BBVA’s organizational boundary)
- Custom GPTs built for specific banking workflows — internal tools trained on BBVA’s own documentation, processes, and terminology rather than generic prompting
- Developer tooling — engineering teams are using the OpenAI API to build customer-facing features, not just internal productivity tools
- Co-development of new AI banking products — the partnership includes joint product work, which is qualitatively different from a typical enterprise software agreement
- AI training and upskilling programs — structured internal education so employees actually know how to use these tools effectively, not just have access to them
The internal upskilling piece is quietly one of the most important parts. A recurring failure mode in enterprise AI rollouts is that companies pay for seats nobody uses effectively. BBVA appears to have invested seriously in adoption, not just access.
The Compliance Question Everyone Asks
Banking is one of the most heavily regulated industries on earth, and AI deployments in financial services carry specific risks around explainability, bias, and customer data handling. How does a 100,000-seat ChatGPT deployment actually work inside that regulatory environment?
ChatGPT Enterprise addresses some of this by design — no training on customer data, stronger admin controls, audit logging. But the harder work is on BBVA’s side: defining which use cases are appropriate for AI assistance, which decisions still require human judgment, and how to document AI-assisted processes for regulators. BBVA operates under oversight from Spain’s CNMV and Banco de España, plus local regulators in every country it operates in. Getting this deployment compliant wasn’t a checkbox exercise.
The fact that they’ve done it, and done it at this scale, is actually useful intelligence for every other bank watching from the sidelines.
What This Means for the Financial Industry — and for OpenAI
Let’s be direct: BBVA is now one of the most visible proof points that large-scale AI deployment in a regulated financial institution is operationally achievable. That matters enormously for the dozens of major banks currently running smaller pilots and trying to build internal business cases for broader rollout.
For OpenAI, this is about more than a big customer. Financial services is one of the highest-value verticals for enterprise AI — banks have complex data, highly educated workforces who can use sophisticated tools, and real willingness to pay for productivity gains. Landing BBVA as a genuine strategic partner, not just a license customer, gives OpenAI a reference client that can credibly speak to other global banks. I wouldn’t be surprised if this partnership ends up being cited in twenty other bank procurement discussions over the next 18 months.
It also puts pressure on OpenAI’s competitors in the enterprise AI space. Google’s Vertex AI and Gemini for Workspace are actively pitching financial services clients. Microsoft Copilot, powered by OpenAI models but packaged differently, is already embedded in many banks through existing Microsoft 365 relationships. Anthropic’s Claude has been making quiet inroads with financial services firms that want a model with a reputation for caution and careful output. The enterprise AI competition in banking is real and intensifying.
BBVA choosing to build its partnership directly with OpenAI — rather than through Microsoft’s Azure OpenAI channel — is a notable signal. It suggests they wanted a closer relationship with the model provider itself, not a reseller relationship. That has implications for how much influence they have over product direction.
Who Else Is Moving Like This?
BBVA isn’t entirely alone. Morgan Stanley has been running one of the most well-documented AI assistant deployments in finance, using OpenAI models to help financial advisors surface relevant research. JPMorgan Chase built its own LLM-based tool called IndexGPT and has been aggressive about internal AI deployment. Goldman Sachs has experimented with generative AI for code generation and document analysis.
But the BBVA announcement stands out for the combination of scale (100,000 employees), geographic spread (25+ countries with different regulatory regimes), and the explicit co-development partnership component. Most bank AI deployments are still primarily productivity tools for internal staff. BBVA is building toward customer-facing AI banking products, which is a meaningfully more ambitious goal.
This connects to a broader trend worth watching: the shift from AI as an internal efficiency tool toward AI as a product surface. As we’ve covered in our look at OpenAI’s broader commercial strategy, the company is increasingly positioning itself as an infrastructure partner for entire industries, not just a productivity tool vendor. BBVA fits that ambition precisely.
The Risk That Doesn’t Get Talked About Enough
Here’s the thing: a 100,000-person deployment is also a 100,000-person dependency. If OpenAI has a major outage, a significant model regression, or a pricing change that makes the economics uncomfortable, BBVA has built significant operational reliance on a single external vendor. Banking regulators increasingly scrutinize third-party technology concentration risk. The questions around OpenAI’s long-term structure and stability aren’t entirely irrelevant here.
That’s not a reason not to do this — every major bank runs on third-party infrastructure in ways that create concentration risk. But it’s a real consideration that any serious risk officer at BBVA is presumably thinking hard about, and it’s one reason the strategic partnership framing matters: a deeper relationship theoretically provides more stability and advance notice of changes than a pure software subscription would.
Key Takeaways
- BBVA has deployed ChatGPT Enterprise to all 100,000 employees globally, making it one of the largest single-organization AI deployments in financial services
- The deal goes beyond a software license — it includes a strategic co-development partnership for building new AI-native banking products
- BBVA’s two-year ramp to full deployment offers a realistic timeline reference for other banks evaluating similar moves
- The deployment includes custom GPTs, API-based product development, and structured employee upskilling — not just chat access
- This puts competitive pressure on Google, Microsoft, and Anthropic all competing for enterprise financial services AI contracts
- Vendor concentration risk is real and worth watching as BBVA’s operational dependence on OpenAI deepens
Frequently Asked Questions
What is ChatGPT Enterprise and how does it differ from the regular ChatGPT?
ChatGPT Enterprise is OpenAI’s business-tier offering with stronger security controls, guaranteed data privacy (conversations aren’t used to train models), higher usage limits, and admin management tools. It’s designed for organizations that need compliance-friendly AI access at scale, which is exactly what a regulated bank requires.
Is BBVA using AI to make customer-facing decisions like loan approvals?
The current deployment is primarily focused on employee productivity tools and internal workflows, with co-development of customer-facing products underway. Automated lending decisions involve separate regulatory requirements around explainability and bias that go beyond what a general-purpose language model deployment addresses — those use cases require more specialized AI systems with different governance frameworks.
How does this compare to what other major banks are doing with AI?
Morgan Stanley and JPMorgan are the closest comparisons in terms of seriousness of deployment, but BBVA’s 100,000-seat scale across 25+ countries and direct co-development partnership with OpenAI puts it at the more ambitious end of the spectrum. Most banks are still in earlier-stage pilots or have more limited internal deployments.
What does the OpenAI partnership mean beyond just using their software?
The strategic partnership includes joint product development work, meaning BBVA and OpenAI are building new banking features together rather than BBVA simply customizing an off-the-shelf product. That’s a qualitatively different relationship that typically involves shared roadmap input, dedicated support, and potentially custom model work tailored to banking use cases.
If this deployment performs as intended, BBVA will have a genuine first-mover advantage in AI-native banking products that competitors will spend years trying to replicate. The financial industry has historically been slow to adopt new technology and fast to copy whoever gets it right first — and right now, every major bank’s digital team is watching what BBVA does next very closely.
