Download the TDB – Official GAII Press Release here
The TDB Global AI Innovation Awards 2026 honoured leading financial institutions that are embedding AI into production systems to transform banking, insurance, and transaction services. HSBC, Citi, JPMorgan Chase, Scotiabank, BMO, Deutsche Bank, TP, Nubank, Peak3, and Bradesco were among the institutions most recognised. With 243 submissions from 91 organisations, the awards revealed that generative AI now dominates the sector, operational improvements are the most reported outcomes, and human oversight remains central as institutions balance automation with governance.
The Digital Banker proudly hosted the TDB Global AI Innovation Awards 2026, recognising the most transformative institutions driving artificial intelligence innovation across global financial services. This year’s winners demonstrated how AI is reshaping banking, insurance, and transaction services, setting new benchmarks for customer experience, operational efficiency, and market leadership.
HSBC stood out with initiatives spanning lending, frontline transformation, wealth management, risk, compliance, and reporting. Its Originations Simplified platform streamlines lending origination by automating document intelligence and workflow processes, doubling straight-through processing and halving turnaround times. It’s agentic AI ecosystem, Wealth Intelligence, provides governed, real-time investment intelligence across seven markets, while the ASK Digital Assistant has reduced risk and compliance advisory turnaround times by more than a third. HSBC also improved frontline knowledge discovery with its AI Knowledge Hub, enhanced reporting through AI-enabled data models, and strengthened fraud resilience with initiatives such as its AML/CFT solution in Hong Kong and its agentic self-learning fraud detection capability. Protection Shield, another Hong Kong innovation, strengthens scam prevention by detecting anomalies and blocking high-risk transactions in real time.
Citi was recognised for embedding AI into workforce capability, sales enablement and wealth advisory. Wingman, a generative AI-powered sales enablement platform, provides Citi’s Home Depot field team with real-time store-level intelligence, replacing hours of manual preparation and strengthening retail acquisition. Citi Advisor Insights supports more proactive wealth advisorythrough predictive analytics and personalised engagement, contributing to significant increases in client assets and revenue. AskWealth, Citi’s first in-house generative AI knowledge assistant, enables wealth teams to retrieve contextual answers in seconds, saving thousands of hours and achieving high user satisfaction. Meanwhile, the Global Workforce Optimization platform supports workforce planning for 40,000 employees, delivering millions in business benefits and laying the groundwork for agentic workforce optimisation.
JPMorgan Chase was recognised for its enterprise-grade AI platforms deployed across global operations. Its proprietary LLM Suite provides secure, enterprise-wide access to AI tools, scaling to 320,000 users across 50 countries and saving hundreds of thousands of hours weekly. Turbo, its AI-powered agentic testing platform, has improved payment assurance by significantly reducing test authoring and regression testing times, while orchestrating the full acceptance testing lifecycle with improved resilience.
Scotiabank was honoured for innovations that enhance customer experience and investment research. Its Predictive Payment Prompts feature anticipates customer needs within the mobile app, achieving high engagement and completion rates for transactions. The Robin AI Platform has streamlined investment research by reducing workflows from hours to minutes, logging thousands of analyst interactions in a single quarter, and enabling faster delivery of market intelligence. Scotiabank also introduced its Pricing Intelligence Platform, a governed multi-model engine that unifies segmentation, elasticity and profitability optimisation, delivering personalised pricing decisions that drive revenue uplift and stronger margins.
BMO was recognised for innovations in lending, mortgage pricing and frontline support. Its Internal Mortgage Renewal Pricing solution generates personalised recommendations that are expected to deliver millions in annualised revenue. Lumi, BMO’s enterprise AI assistant, has reduced support call volumes by more than half, and expanded to support thousands of frontline employees with bilingual conversational access to trusted policies and procedures. BMO also introduced MINT, its AI-powered credit loss forecasting ecosystem, which integrates dozens of predictive sub-models to forecast default and insolvency outcomes up to 18 months ahead.
BNP Paribas was honoured for advancing AI-driven banking through innovations that improve customer support, personalisation and operational efficiency. Its Genius Bar virtual assistant and Hello bank!, has reduced negative chatbot feedback to under 1% and doubled self-service resolution rates. Meanwhile, its AI-powered mortgage solutions support 80% of mortgage originations, saving time for relationship managers and cutting overall processing times by 10 days.
TP.ai Banking was recognised for creating proactive, agentic AI-powered customer experiences across collections, compliance and dispute resolution. Serving hundreds of global clients, the platform has delivered significant reductions in cost to collect, improvements in recovery rates, and efficiency gains across compliance and chargeback processes, all while maintaining human oversight and embedded guardrails.
Nubank was named World’s Best AI-Powered Digital Bank forbuilding one of the earliest AI-native bank at scale with 135 million customers across Brazil, Mexico, and Colombia. Its proprietary foundation model, nuFormer, driving record credit card market-share gains. AI agents now manage millions of customer interactions with human-parity satisfaction, while the AI Private Banker delivers personalised financial guidance to over 15 million users.
Peak3 secured the World’s Best AI-Powered Insurance Technology Provider award with its AI-Powered Document & Voice Automation Suite. The solution automates data extraction across claims, underwriting, motor, and health documents with 95% accuracy, reducing processing times by 93%. Voice-enabled first notice of loss shortens claims intake from 15 minutes to under three, while AI-generated quotations cut preparation from 20 minutes to two.
Bradesco was awarded World’s Best AI-Powered Retail Bank for its enterprise AI platform, BRIDGE. Supporting hundreds of use cases across customer experience, operations, risk management, and technology, BRIDGE integrates generative AI, agentic AI, governance, and multi-cloud infrastructure to enable AI deployment at scale. By embedding AI into core operations, Bradesco has improved productivity, accelerated innovation, strengthened customer engagement, and enhanced decision-making.
Important findings from this year’s submissions
This year, the TDB AI Innovation Awards received 243 submissions from 91 organisations, covering 179 distinct initiatives. From these submissions our research indicates the following:
Generative AI dominates the field. More than two-thirds of the initiatives featured generative or hybrid AI approaches, with 40% purely generative and 29% hybrid. Predictive machine learning accounted for 22%, while 9% did not specify their approach. Generative AI is primarily being deployed internally, with nearly three in four initiatives used as staff assistants rather than customer-facing tools.
AI adoption starts with staff workflows. Half of the submissions were designed exclusively for internal use, while only 24% were customer-facing. The remaining 26% combined staff and customer components, underscoring how institutions are embedding AI into their own operations before extending it outward.
Production use is now the norm. Nine in ten submissions described AI in production rather than pilot stages. Of the 160 initiatives analysed, 144 were already live, with 119 reporting every described application in production and 25 presenting portfolios with at least one live component.
Human oversight remains central. In 58% of initiatives, AI assists or recommends while a person makes the final decision. Hybrid systems were the most action-oriented, with 68% executing automatically or after approval, compared with 23% of generative and 29% of predictive initiatives. Wealth management stood out as a domain where human advisers remain firmly in control, with no automatic execution reported.
Safeguards are widely described but unevenly validated. Nine in ten initiatives reported some form of safeguard, with human oversight the most common at 41%. Yet fewer than one in eight reported a validation result, highlighting a gap between safeguard design and demonstrated effectiveness.
Operational improvement is the leading business outcome. Forty-one percent of initiatives reported measurable operational gains, compared with 24% citing customer results, 20% financial outcomes, and 6% risk improvements. However, reporting remains uneven: while 71% provided a production business figure, only 23% included a baseline, period, and population for the same metric.
Technology transparency is limited. Few institutions disclosed which large language models they used, with only 21% naming the model family. Meanwhile, “agentic” terminology was common, appearing in 38% of submissions, but often referred to assistance or recommendation rather than autonomy.
Together, these findings reveal a financial services industry moving decisively beyond experimentation. AI is being embedded into production systems, reshaping staff workflows, and delivering measurable operational improvements, while institutions continue to balance automation with human oversight and safeguard design.
