The Bank of England Retail Payments Infrastructure Board (RPIB) consultation on the Design of the Future Retail Payments Infrastructure for the United Kingdom makes for interesting reading, for people who are interested in this sort of thing. I started to read through it and came across the diagram at the heart of the consultation
I was looking forward to seeing some next generation thinking but my eyes were immediately drawn to the bottom of the “stack” where the only two possibilities were set out for the underlying payment infrastructure intended to support the new economy. These were gross settlement and net settlement, exactly as we have now. In fact, exactly as we have had for a generation. Conspicuous by its absence was the idea of no settlement. In other words, instead of taking the existing structure whereby there is a messaging layer that clears transactions between institutions sitting on top of a settlement layer that moves the value between those institutions, a more futuristic view of the infrastructure, congruent with the Bank for International Settlements (BIS) vision of next-generation Financial Market Infrastructure (FMI). In the BIS view, the “Finterent” would be based on some form of shared ledger solution whereby there is no distinction between the message and the money.
In such a model, to paraphrase McLuhan, the money is the message.
I must stress here that this is not solely a technologists’ perspective. I have consistently said that tokenisation will be driven by money people not technology people and far more important than my views are the views of people who understand the big picture on financial services. To take just one such person, Sir Jon Cunliffe (a former Deputy Governor for Financial Stability at the Bank of England and a former Chair of the Bank for International Settlements Committee on Payments and Market Infrastructures) used an interesting analogy. He compared the payment systems we use today to the piston and propellor engines reaching the ceiling on performance just as jet engines (ie, tokenisation) were being developed. In his view, the development and deployment of digital assets together with the code that governs their use, transfer and ownership (ie, the world of decentralised finance, or DeFi) means a quantum leap in payments. And for what it’s worth, I could not agree more.
In that tokenised world, markets trade digital assets that act as a proxy for whole or fractional ownership of real-world assets. Tokens will massively benefit markets by enabling illiquid asset fractionalization by programmatically enforcing the required rules and restrictions, allowing for enhanced mobility of asset ownership, efficient payments and distributions, and a range of financial benefits across the lifecycle of the transactions.
Given that Sir Jon has undoubtedly forgotten more about financial services than I will ever learn, I am reluctant to criticise but I think he may not have been hyping the sector enough! In 2024, Standard Chartered estimated that the market size of real-world tokenised assets will climb as high as $30 trillion by 2034. Right now, the value of tokenized real-world assets on public blockchains climbed from roughly $5.8 billion in early 2025 to more than $30 billion by April 2026 — a jump of over 420% in about sixteen months, according to data tracked by RWA.xyz. Include stablecoins, and the broader tokenized market sits north of $240 billion.
Sarah Breeden, a current Deputy Governor at the Bank of England, has also observed that tokenisation could allow “payments to be embedded more efficiently and deeply into our increasingly digital economy” by which I am sure that she meant that allowing the direct exchange of value (ie, the exchange of fungible and non-fungible tokens through automated means) would radically reduce both costsand risks, since all exchanges would be between bearer instruments. Hence my point that the drive for tokenisation in financial markets has nothing to do with ideology or libertarian appeals for decentralisation. It is about money. It’s about replacing intermediaries with apps. It’s about trading and transacting in a wallet-to-wallet world.
Taking all of this on board, then, I think that the emphasis in the consultation should be shifted and that the core infrastructure should be explicitly designed to accommodate regulated stablecoins, DeFi and agent‑mediated payments as critical components of the UK payments ecosystem, even if their primary role is “under the hood” rather than directly visible to retail end users. In other words, the core should assume a future in which bank deposits, a potential digital pound and regulated sterling stablecoins are all live and interoperable, and in which programmable, agent‑mediated payments are routine rather than exotic. That is not some distant science fiction projection of one of a number of possible futures, it is the future that is being built right now.
Response to Consultation
## Q1: Future payment journeys – stablecoins, DeFi, identity and multi‑money
The list of proposed payment journeys provides a good baseline, but it under‑specifies the implications of a genuinely “multi‑money” ecosystem that includes regulated stablecoins, DeFi‑based execution and agent‑mediated flows. I would therefore support the current list and recommend the following augmentations.
First, the consultation references “payments across different forms of money” and explicitly mentions regulated stablecoins and a potential digital pound. However, it is not yet clear whether the core is envisaged as a neutral “hub” that can route payments between bank money, stablecoins and a digital pound, or as a system primarily for bank money, with stablecoin interoperability left to bilateral arrangements at the product layer. Given the Bank’s work on systemic sterling stablecoins and the expectation of live arrangements later this decade, the architecture should explicitly assume that systemic stablecoin issuers and wallets may wish to connect to the core as participants in their own right. That has clear consequences for messaging standards, access models and the design of common utilities (e.g. fraud analytics and alias directories) which should be exposed to stablecoin PSPs on a proportionate basis.
Second, the “programmability” and “advanced delegated payments” journeys should be framed in a way that anticipates DeFi‑style composability and agentic commerce. The consultation already anticipates AI agents initiating delegated payments and programmable rules for conditional transfers, which is a natural bridge to on‑chain execution environments and other programmable services. To make this concrete, I suggest explicitly adding journeys such as: (i) programmable escrow and milestone‑based payments triggered by verifiable oracles (e.g. delivery, performance); and (ii) agent‑managed budgets and subscriptions (for households and SMEs), where an AI agent executes small, frequent payments within user‑defined constraints. These journeys will often be implemented using smart‑contract‑like logic, whether on traditional infrastructures or DeFi‑compatible ledgers, and embedding them now in the canonical list makes it easier to justify the modular and extensible architecture that the design principles envisage.
Third, digital identity and agent‑mediated payments should be described more explicitly. The consultation already highlights the importance of common utilities and shared services, including fraud data and alias directories, and references delegated and agent‑initiated payments as a key future journey. I would encourage the Bank and DeliveryCo to treat digital identity as the trust fabric for both human users and AI/agentic systems operating across multiple forms of money.
For human‑centric digital identity, in a multi‑rail, multi‑money environment it will be increasingly important that individuals and businesses can present high‑assurance, reusable digital identities—potentially aligned with eIDAS‑style wallets and verifiable credentials—to support onboarding, KYC and continuous risk management. The core should assume that participants will rely on interoperable, credential‑based identity schemes, allowing users to prove attributes (e.g. name, age, residency, account ownership) selectively rather than re‑onboarding for each provider. These identity proofs should be bindable not only to traditional accounts, but also to stablecoin wallets, tokenised deposit instruments and any future digital pound wallets, providing a consistent trust and liability framework across forms of money. This supports the emphasis on common utilities: identity, fraud analytics and sanctions screening become shared services rather than duplicated, siloed functions.
For identities of AI agents and automated systems, as delegated and agent‑initiated payments become more common—whether as personal finance “copilots” or enterprise agents orchestrating just‑in‑time payments—experience from early deployments suggests that agents require their own distinct digital identities, separate from the human account holder, tied to cryptographic credentials that can be authenticated and revoked independently. Trust in agents depends on being able to verify not only which human they represent, but also which agent runtime, which organisation owns it, and what capabilities and constraints have been delegated. The core conceptual architecture should therefore support a model in which humans, organisations and agents are all first‑class identities, each able to hold and present verifiable credentials binding them to specific roles, permissions and regulatory obligations, and a clear separation between authentication (proving who or what is initiating a payment or instruction) and authorisation/delegation (what that entity is allowed to do, under what constraints).
Finally, cross‑border payments and interlinking arrangements should explicitly contemplate interoperation not just with overseas payment systems, but also with regulated tokenised money and securities platforms that may themselves be DeFi‑based or DeFi‑adjacent. Given the global direction of travel under the G20 cross‑border roadmap and international work on stablecoins and crypto‑assets, it is prudent to treat these as likely counterparties rather than edge cases.
***
## Q2: Migration timelines and product roadmaps – integrating new forms of money and agents
Migration timelines and product roadmaps should reflect that the introduction of systemic stablecoins, potential digital pound arrangements and agent‑mediated payments are not independent events but deeply intertwined with core infrastructure renewal. There are at least three relevant considerations.
First, the roadmap envisages staged design, build and migration of the next‑generation infrastructure, in coordination with Pay.UK, while other workstreams (stablecoins, digital pound, digital identity) are advancing in parallel. It would be helpful to synchronise these workstreams explicitly, for example by ensuring that DeliveryCo’s initial builds and prototypes include at least one multi‑money use case (such as account‑to‑stablecoin payments) and at least one agent‑mediated use case (such as delegated bill management) as non‑negotiable requirements. Stablecoin and “new money” interoperation, and agent‑mediated flows, should be early deliverables, not phase‑two enhancements.
Second, transition planning should anticipate that some PSPs may wish to connect via multiple rails simultaneously—RTGS, the new retail core and one or more regulated stablecoin networks—while also deploying agentic front‑ends. This raises practical questions about intraday liquidity, queue management and customer protection across these environments during dual‑running, and about expectations on participants who offer customers both bank accounts and stablecoin wallets managed by agents. Outcome‑level ambitions around seamless multi‑money payments cannot be achieved if such questions are left entirely to the product layer.
Third, the roadmap for programmability and agentic behaviour should be explicitly tiered. Tier 1 could focus on basic programmable features (scheduled rules, simple conditions) and human‑initiated delegated payments with tight limits, while Tier 2 introduces more sophisticated conditional logic and AI‑driven agents, subject to default caps, revocation mechanisms and clear attribution of initiation methods. This staged approach aligns with the design principle that the core should be modular and extensible, enabling new features without wholesale redesign.
***
## Q3: Design principles – multi‑ledger interoperability, on‑chain risk and identity
I broadly agree with the proposed design principles and characteristics, but recommend two additions and one clarification.
First, under “platform for innovation and competition,” the principle of multi‑ledger interoperability should be made explicit. The consultation rightly emphasises modularity and extensibility, but does not explicitly state that the core will need to interact with multiple ledgers (RTGS, the new retail core, stablecoin ledgers, potentially a digital pound infrastructure). A design principle such as “supporting safe interoperability across multiple settlement and recording infrastructures” would justify investment in canonical interfaces, messaging patterns and risk controls for bridges and gateways.
Second, under “security and resilience,” the threat model should explicitly reference on‑chain risk and bridge risk. DeFi‑style environments—whether internal or external—introduce new classes of vulnerabilities (smart‑contract bugs, oracle manipulation, bridge exploits) that can propagate back into the traditional system if not properly bounded. Explicit recognition of these risks will encourage segregation and ring‑fencing of on‑chain components, layered controls and monitoring for multi‑rail routes, and proportionate resilience requirements for intermediating entities (trust anchors, tokenisation platforms).
Third, the design principles should acknowledge identity and delegation as first‑class concerns. Consent and delegation should not be treated as static flags but as programmable objects with lifecycle and provenance. Scheme rules and infrastructure should define standardised ways to express delegated payment mandates (what agent may spend, on which instruments, within which limits and for how long), require that every payment and programmable trigger carries metadata about who authorised it and which agent (if any) initiated it, and provide standard revocation and pause mechanisms so that consumers and businesses can rapidly withdraw consent from compromised or misbehaving agents without closing underlying accounts.
***
## Q4: Trade‑offs and prioritisation
The main trade‑offs arise between performance and scalability, security and resilience, and the ambition to be an open, programmable platform for innovation. Adding stablecoins, DeFi‑style programmability and agentic flows sharpens these tensions but does not fundamentally change them.
From a performance perspective, it might be tempting to treat complex programmable flows and multi‑money routes as edge cases and push them entirely out to the product layer. However, doing so risks fragmentation and inconsistent standards, undermining competition and user trust. A more balanced approach is to keep the core state lean, focusing on clearing and finality, while providing standardised hooks and data models for external execution environments, including DeFi platforms and smart‑contract infrastructures. The core should be the backbone, not the brain, but a backbone designed for a programmable, multi‑money, agent‑mediated future rather than a slightly upgraded version of today’s RTGS/FPS world.
On resilience, there will be cost and complexity associated with multi‑rail, multi‑money interoperation and identity‑rich agentic flows, especially where on‑chain environments are involved. Proportionality and risk‑based access are therefore key: systemic stablecoins and critical DeFi‑style services should be held to higher resilience and governance standards than niche offerings, while smaller, innovative firms should not face insurmountable barriers to connecting.
Finally, there is a trade‑off between top‑down control and open, modular evolution of the ecosystem. The layered conceptual architecture—scheme/standards, core infrastructure and services—offers a natural way to balance these pressures. Extending the services layer to include programmable payment modules, multi‑money conversion services and identity/agent utilities will require careful governance, but it offers a route to leverage global innovation without diluting UK public policy objectives.
Snippings 