Inside the Bank Where Employees Built 20,000 AI Tools in 18 Months

Key Takeaways

• BBVA employees created 20,000 custom GPTs in 18 months
• 80% of 120,000 staff use the platform daily
• Peru assistant cut query time by 87%
• Scarcity of 3,000 licences drove high adoption
• Peer‑led "Champions and Wizards" network replaced formal training

Pulse Analysis

The financial sector has poured roughly $252 billion into AI, yet most firms struggle to extract measurable value. BBVA’s experience flips the script by treating employee‑initiated "shadow AI" as a demand signal rather than a compliance breach. By providing a safe, governed platform, the bank turned a potential risk into a prolific source of custom solutions, demonstrating that scale does not require top‑down control alone.

Key to BBVA’s success was a deliberate scarcity strategy: only 3,000 licences were initially released for a 125,000‑person organization. This artificial constraint created urgency and fostered a sense of exclusivity, prompting rapid peer‑to‑peer diffusion. The "Champions and Wizards" network—300 internal advocates without a formal curriculum—served as organic trainers, driving the platform’s 80%+ daily active usage and ensuring that tools aligned with real‑world frontline needs.

The outcomes extend beyond headline metrics. A Peru‑focused operations assistant slashed query response times by 87%, while a legal GPT now handles 40,000 annual inquiries with a nine‑person team, freeing senior lawyers for higher‑value work. Moreover, language‑specific translators opened BBVA’s products to Nahuatl and Maya speakers, illustrating how employee‑built AI can unlock underserved markets. For banks and other regulated enterprises, BBVA’s case offers a replicable blueprint: combine limited access, peer empowerment, and robust governance to convert shadow AI into a strategic asset.