AI in Banks

Česká spořitelna's Future Lab branch in Prague runs the Czech Republic's first 24/7 AI-powered bank lobby, where a Promethist Engage digital human greets every visitor in 20+ languages, qualifies intent, and summons bankers with full context. Innovations from the lab are now feeding into Česká spořitelna's 320 branches nationwide and across Erste Group's seven-country network.

ING and McKinsey & Company deployed a generative AI chatbot for banking customer support in the Netherlands, which handles ~85,000 customer interactions weekly. Built in seven weeks and launched in September 2023, it helped 20 percent more customers resolve issues in its first seven weeks than the previous rule-based bot.

Hong Kong-based fintech Apoidea Group, creator of the SuperAcc document-processing platform, worked with AWS to fine-tune the Qwen2-VL-7B-Instruct vision-language model with LLaMA-Factory on Amazon SageMaker HyperPod. The fine-tuned model raised TEDS table-structure-recognition accuracy from 23.4 to 81.1, beating Anthropic Claude 3 Haiku on the FinTabNet benchmark and enabling SuperAcc to cut financial-spreading review from 4–6 hours to roughly 30 minutes for its 10+ FSI clients.

ING Belgium's Site Reliability Engineering team faced severe class imbalance when training a machine learning model to predict whether planned IT changes would lead to Major, P1, or P2 incidents. By augmenting training data with synthetic replicas generated through DataCebo's Synthetic Data Vault (SDV) Enterprise, the team improved a LightGBM classifier's F1 score from 76% to 97% — a 21 percentage-point gain. The deployment covers hundreds of daily IT changes logged in ServiceNow and Jira Service Management.

Banco Santander announced June 22, 2026 that it has deployed 280 production AI agents, hit $40M ROI in Q1 2026, and extended AI access to all 185,000 employees worldwide; targeting $1.15B in business value by 2028.

Banca Sella, a 140-year-old Italian banking group serving 1.4+ million customers, deployed Prometheux for credit risk analysis, transforming monthly credit score explanations from 17 minutes to 5 seconds.

Citi's Services unit deployed four generative AI tools — Sales Assist, Ops Assist, RFP response automation, and Client Assist chatbot on Citi Direct — cutting document processing time by approximately 80% (from hours to 10-15 minutes per client onboarding case), with Client Assist currently in pilot.

China Merchants Bank (CMB) deployed DeepSeek-V4 Flash large-EP inference on Huawei Ascend NPU: 1 Router + 2 Prefill + 2 Decode topology, each inference instance spanning 16 NPUs (1 Leader + 15 Workers) — 64 inference NPUs + 1 Router total. SGLang RBG provides dynamic port allocation, three-layer service discovery, in-place upgrade, and instance-level self-healing. Industry's first DeepSeek-V4 inference rollout on a domestic AI chip stack with cloud-native orchestration.