By Miguel Santos Luparelli Mathieu, Product Innovation Director
While the first part of the 2026 Radar focused on infrastructure, this second part shifts the spotlight to how and where trust decisions are made. Intelligence is progressively moving towards the device, the context and user behaviour, redefining how digital security is built.
In this second part, we explore how these advanced AI capabilities will transform the design of financial applications, regulatory compliance, and the relationship between people, devices and intelligent systems.
Physical AI and Decentralisation: Intelligence Moves to the Device
Physical AI—endowing devices with the ability to perceive context and environment—combined with the decentralisation of data collection, processing and decision-making at the device level, will shape the design of financial applications in 2026.
Concepts such as Device Intelligence, Contextual Awareness and Passive Behavioural Biometrics will consolidate as foundational signal blocks that Physical AI will use to continuously assess identity throughout enrolment, authentication and transactions.
Wallets can act as:
- an external trust layer in standalone, industry-specific applications
- or an internal source of intelligence embedded within financial applications
In addition, on-device Agentic AI is expected to accompany users throughout their entire journey—guiding processes, strengthening security and offering recommendations without adding friction.
Key impact
- Behavioural Biometrics
- Device and contextual intelligence
- Agentic AI
AI-Native Systems and Adaptive User Experience
AI-native applications—designed from the ground up with AI as a core capability—and AI Agents built for specific tasks and integrated into Agentic AI for autonomous decision-making, were the main innovation drivers in 2025 and will define digital experiences in 2026.
Application design will become simpler:
- fewer screens
- less friction
- more intuitive interactions
AI will support document verification, recommend the use of verifiable credentials, validate identity attributes and generate automated narratives for SARs (Suspicious Account Reports) and STRs (Suspicious Transaction Reports), reinforcing compliance while reducing abandonment rates.
Key impact
- Identity Verification and Authentication
- Agentic AI
- KYC, pKYC and screening
- Automation of SARs and STRs
Bonus Track: Beyond 2026
World Models: Two Approaches to Artificial Intelligence
The future of AI is currently shaped by two competing approaches.
On one side, Large Language Models (LLMs), primarily trained on language, stand out for their immediate responsiveness. Following Daniel Kahneman’s framework, they represent “fast thinking”.
On the other, researchers such as Yan LeCun advocate for the development of World Models—systems that learn from video and spatial data to understand the world. This paradigm shift would enable machines to achieve “human-level intelligence”, understand reality, and, according to LeCun’s vision, represent the true path towards Artificial General Intelligence (AGI).
Impact
- KYC personalisation
- Automation of SARs and STRs
- Adaptive pKYC
Quantum Analytics
Quantum analytics opens up new possibilities for fraud detection by leveraging principles such as superposition and entanglement—particularly in models based on categorical data and complex relationships that are difficult to uncover with traditional computing.
For example, in risk modelling, a financial institution aiming to detect transaction fraud could use quantum algorithms to pre-process data and uncover hidden relationships between features, enabling traditional computing to identify fraud patterns more effectively.
Impact
- Transaction Monitoring
- Account Classification (Know Your Account)
Energy Consumption for AI
AI energy consumption—especially for Agentic AI—will become a strategic factor. Nuclear power is shaping the future of energy supply required to fuel both the training and inference of Generative AI and its most impactful derivatives: Agents. Domain-Specific Language Models will play a key role in improving efficiency and scalability.
Impact
- Agentic AI
The future of digital identity will be contextual, continuous and deeply embedded in the user experience.
Beyond 2026, competitive advantage will not come from adding more controls, but from making better trust decisions with less user intervention. Organisations that invest in intelligent, ethical and energy-efficient identity systems will be better prepared to face the next decade of fraud, regulation and the digital economy.