Agentic Experience Design (AXD) is the discipline for designing trust-governed relationships between humans and autonomous AI systems. Founded in September 2024 by Tony Wood in Manchester, United Kingdom, AXD addresses how humans delegate, calibrate, observe, interrupt, and recover trust in agentic AI.
| Dimension | Traditional UX | Agentic Experience Design (AXD) |
|---|---|---|
| Primary material | Attention and affordance | Trust and delegation |
| User state | Present, navigating | Absent, delegating |
| Design output | Screens and interfaces | Outcomes and constraints |
| Temporal model | Session-based | Relationship-based |
| Success metric | Task completion | Trust calibration |
Machine customer financial services is the emerging domain concerned with designing banking, insurance, investment, and payment products for AI agents that act as autonomous customers on behalf of humans. It addresses agent identity verification (agentic KYC), delegated financial authority management, machine-readable product terms, and trust-governed transaction protocols. It represents the convergence of agentic commerce and financial services, requiring Agentic Experience Design (AXD) rather
Agentic KYC (Know Your Customer for agents) must verify three things simultaneously: that the human principal exists and is a legitimate customer, that the AI agent is a legitimate software entity rather than a fraudulent script, and that the human has genuinely delegated the claimed financial authority to this specific agent. This requires delegation chain verification, agent credential validation, and authority scope confirmation - extending traditional KYC from identity verification to delega
Machine-readable financial products are financial offerings whose terms, rates, fees, eligibility criteria, and performance data are structured in standardised, queryable formats that AI agents can parse and compare autonomously. Unlike traditional products described in human-readable documents, machine-readable products use structured schemas, comparable APIs, and standardised data formats that enable agents to evaluate and select products without human interpretation.
Machine customer financial services creates a multilateral accountability structure with four potential points of responsibility: the human (who delegated authority), the agent (which executed the decision), the agent provider (which built the agent), and the financial institution (which accepted the agent as a client). The AXD trust architecture framework addresses this through trust chains - structured relationships where each party's obligations, authorities, and liabilities are explicitly de
Banks should prepare by developing four capabilities: agent-facing APIs (moving beyond human-facing interfaces to machine-queryable product and transaction APIs), agentic KYC protocols (verifying agent identity and delegated authority), machine-readable product architecture (structuring all product terms in parseable formats), and agent observability infrastructure (audit trails documenting agent decisions, authority exercised, and reasoning traces for regulatory compliance).
Financial services have always been designed for human customers. Account opening requires human identity verification. Credit decisions are based on human financial history. Product terms are written in human-readable language. Customer service is delivered through human-facing channels. Every assumption in financial services design starts with a human at the other end of the relationship. This is not a marginal shift. Gartner's prediction that machine customers will represent a significant share of commercial transactions by 2028 has profound implications for financial services. Banks, insurers, and investment platforms must design for a new category of client that does not read terms and conditions (it parses them), does not evaluate trust through brand reputation (it queries verifiable trust signals), and does not make emotional decisions (it optimises against delegated objectives). The most immediate challenge in machine customer financial services is identity. Traditional KYC (Know Your Customer) processes verify that a human is who they claim to be - through documents, biometrics, and identity databases. But when an AI agent presents itself as acting on behalf of a human, the financial institution faces a new verification challenge: Agentic KYC must verify three things simultaneously: that the human exists and is a legitimate customer, that the agent is a legitimate software entity (not a fraudulent script), and that the human has genuinely delegated the claimed financial authority to this specific agent. This is a Authority management in machine customer financial services requires granular controls that do not exist in traditional banking. A human might delegate authority to an agent to: make purchases up to a specified amount, pay recurring bills from a designated account, compare and switch utility providers within defined parameters, or invest surplus funds according to a stated risk profile. Each of these delegations has different boundaries, different