By Leonora Smart-Abbey*
SUVA, Fiji | 23 September 2026 (IDN) — Digital identity, digital payments and data exchange have become powerful reference points for thinking about what digital public infrastructure (DPI) can enable. They have supported financial inclusion, improved access to services, enabled new forms of digital participation and demonstrated how common digital foundations can create value across sectors. However, as emerging technologies advance, digital transformation is moving into a new phase.
Artificial intelligence is changing how systems process information and make decisions. New forms of digital credentials and trust are emerging, data architectures are becoming more distributed, and transactions are becoming increasingly programmable. Digital systems are interacting across institutional, sectoral and national boundaries in ways that were less prominent when many of today’s DPI concepts took shape.
With this evolution, it is highly important to review the conceptualisation of DPI against the new technologies. The current framing has been relevant to advocacy and initial support for DPI; however, its outcome-focused conceptualisation risks becoming obsolete as technologies evolve. The building blocks (BBs) of DPI should be redefined around their underlying functions, which will remain necessary regardless of how emerging technologies evolve.
From Today’s Digital Tools to Tomorrow’s Infrastructure
Each generation of digital transformation tends to organise itself around the technologies and applications most visible at the time. This made sense: governments needed tangible systems, development partners needed practical investment areas, and citizens needed services that solved immediate problems, but visibility does not necessarily make something foundational.
Digital identity, for example, is an enormously important capability. Yet the identity a person uses digitally is not created by a single technological component. It depends on authoritative information about that person, mechanisms that authenticate and authorise interactions, credentials and trust services, and the ability of relevant systems to exchange information.
The same is true of digital payments. A payment is an outcome of underlying functions that establish who or what is transacting, what can be authorised, how the transaction is executed, and how relevant information moves between systems. Payment systems are therefore critical, but they result from a combination of registries, exchange, and transactional or mediating capacities within a broader digital architecture.
This suggests that because both key DPI BBs depend on similar registry and system-mediator capacities, DPI systems developed in isolation may duplicate effort.
For instance, a national personnel identification system and an electoral identification system may be developed to meet different institutional needs and serve different populations or use cases. The personnel system may be designed to establish and verify an individual’s identity for access to government employment and related public services. In contrast, the electoral system may be designed to establish voter eligibility and authenticate citizens for participation in elections. On the surface, these appear to be distinct digital identity solutions serving distinct policy objectives. Yet beneath these different applications are potentially overlapping functions: establishing authoritative information about an individual, issuing or validating credentials, authenticating identity, determining eligibility, and authorising access to a particular service or transaction. If each system develops these functions independently, government may effectively create multiple identity registries, credentialing mechanisms and authentication or authorisation services that perform similar functions but operate within separate institutional and technical silos.
Additionally, the populations covered by such systems can substantially overlap. Indeed, electoral populations are generally a subset of the broader population that national identification systems seek to establish and verify. Nevertheless, the immediacy of electoral requirements can mean that electoral identification infrastructure is developed before, or independently of, a comprehensive national identification system, which is a common case in developing countries. When a national ID system is subsequently introduced, it may be designed as a separate program rather than as an extension of, or interoperable layer within, the existing identity ecosystem.
This is not necessarily the result of poor planning. Different systems often emerge from different institutional mandates, procurement processes, political priorities, funding arrangements and implementation timelines. Each program can therefore be rational within its own context while collectively contributing to a fragmented identity architecture.
As a result, governments may end up maintaining multiple registries, credentialing mechanisms, authentication processes, and authorisation arrangements that perform overlapping foundational functions. The issue, therefore, is not that different sectors have different identity needs – this is acknowledged. The issue is whether those needs require entirely separate foundational infrastructure, or whether shared, interoperable functions can support them while allowing applications to retain their distinct mandates, governance arrangements, and safeguards, while drawing on common foundations.
This distinction matters because it shifts the question from “How do we build a digital identity system for this particular use case?” to “What foundational identity-related functions need to exist, and how can different use cases securely and appropriately access them?“
The above example shows duplication within a BB: two systems both fall under digital identity, yet may independently reproduce registries, credentialing, authentication, and authorisation functions. However, the risk of duplication and fragmentation is not limited to a single building block. It can also emerge across BBs when outcome capabilities conceptually treated as distinct depend on similar underlying functions.
Consider the relationship between digital identity and digital payments. These are commonly treated as separate foundational BBs because they address different needs: identity enables individuals and entities to establish and verify who they are, while payments enable them to transfer value. Yet, at the functional level, both require a range of similar capacities. A digital payment may need to establish the identity of the transacting parties, authenticate the actor initiating the transaction, determine whether that actor is authorised to perform it, validate relevant credentials, apply rules governing the transaction, and create an auditable record of the action. A digital identity system similarly requires mechanisms for credentialing, authentication, authorisation, and trusted interaction.
This does not mean that identity and payment systems should be merged. Their institutional mandates, regulatory requirements, risk profiles and governance arrangements may appropriately remain distinct. The point is that treating them as entirely separate foundational BBs can obscure the common functions that sit beneath both. If those functions are independently designed and embedded within each system, governments may reproduce similar authentication, credentialing, authorisation, and trust mechanisms across systems intended to form part of a common digital infrastructure.
The issue therefore extends beyond duplication between two systems performing broadly the same function. It concerns the possibility of reproducing the same foundational functions across systems performing different functions. This suggests that the architecture of DPI may be better understood not only through the outcome capabilities it delivers, but through the underlying functions that enable those outcome capabilities to interact. Digital identity and digital payments can remain distinct outcome capabilities while drawing, where appropriate, on common foundational functions for establishing trust, authenticating actors, authorising actions and enabling secure transactions.
A Functional Foundation for DPI
Therefore, the foundational infrastructure of DPI should be defined by the functions it performs, while digital outcome capabilities should be understood by the outcomes those functions enable. This becomes increasingly important as the range of digital outcomes expands. With emerging technologies, future foundational digital systems may need to support not only identity and payments, but also credentials, automated decisions, digital rights, geospatial services, climate information, machine-to-machine interactions, tokenised assets, AI-enabled verification, and entirely new forms of digital trust that have yet to become mainstream. A framework defined around today’s outcome possibilities risks repeated redesign as these emerging technologies evolve, but a framework defined around enduring functions can evolve with them. Once this distinction is made, DPI becomes less tied to a particular generation of technology.
The objective is not to replace the implementation of digital identity or digital payments. It is to understand them more precisely as outcome capabilities within the architecture that enables them and, rather, redefine the BBs of DPI by their functional capabilities.
A more future-ready approach is to conceptualise DPI around three functional BBs, namely Digital Registries, System Mediators, and Data Exchange Mechanisms. These provide a deep architectural lens for understanding existing and emerging capabilities.
- Digital Registries: Establishing what is true
Digital registries provide the foundational data layer. They are authoritative systems that structure, store, and govern core information about people, entities, assets, locations, or other objects relevant to public and economic activity. Population registries, civil registries, business registries, land registries, health registries, education registries and social registries are familiar examples. However, the principle extends beyond traditional administrative data. As digital transformation expands, governments may increasingly need registries for geospatial information, climate assets, infrastructure, digital entities, IoT devices and other forms of information that become important to policy and service delivery. Irrespective of these advancements, the underlying function remains the same: establish authoritative, structured and governable information that other systems can reliably reference. This is why the registry is foundational. It establishes what is true. It does not, by itself, determine what a person or organisation can do with that information. That requires another set of functions.
- System Mediators: Determining what can be done
System mediators are the functional services that act upon information and enable secure digital interactions. They can authenticate, authorise, credential, transact, verify, orchestrate workflows, manage consent, enforce rules, establish trust and record actions. Authentication services, credentialing systems, trust frameworks, workflow engines, authorisation services, consent management, digital signatures, and other transactional services all mediate between data, actors, and applications. This distinction becomes particularly important as new technologies emerge. An AI-enabled verification service, a digital credential, a tokenised transaction or an automated decision mechanism may look very different technologically. Yet each can perform a mediating function within a wider digital system. The enduring revelation, then, is not which technology performs the mediation. It is what function needs to be performed, under what rules, with what safeguards and with what authority, and system mediators answer that.
- Data Exchange Mechanisms: Determining who can access what, how and when
The third building block is the connective layer. Data exchange mechanisms enable information to move between registries, mediators, applications and other systems in ways that are secure, structured, governed and interoperable. Today, this may involve APIs, interoperability platforms, messaging systems and secure data-sharing protocols. Tomorrow, the architecture may involve event-driven systems, decentralised data pathways, semantic interoperability, new forms of cross-border digital exchange or AI-mediated information flows. The technology may change, but the requirement does not. Digital systems need mechanisms for information to move between actors according to defined rules, permissions, standards, and safeguards, and data exchange informs it.
Together, these three functions create a more durable way of understanding DPI:
Registries establish what is true.
Mediators determine what can be done.
Data exchange determines who can access what, how, and when.
Digital identity, digital payments and other digital capabilities can then be understood as outcomes generated through the interaction of these foundational functions, rather than as isolated infrastructure components.
The proposed DPI BBs should not be understood as independent or self-sufficient components. They are functionally distinct, but interdependent layers of infrastructure. The distinction is therefore not between functions that operate independently, but between different functional roles within the system: establishing authoritative state, mediating interaction with that state, and enabling governed exchange of information.
Designing DPI for Technological Evolution
This distinction becomes clearer when we consider emerging technologies. A government should not need to redefine its DPI architecture every time a new identity model emerges, a new payment technology becomes viable, or a new method of data exchange changes how systems interact. A functional architecture should be flexible enough to accommodate such changes. The proposed model posits that:
- A registry can evolve in how information is stored, structured or represented without changing the fundamental requirement for authoritative data.
- A mediator can evolve from conventional authentication and payment services toward new forms of credentialing, programmable transactions, AI-assisted verification or other trust mechanisms without requiring the entire DPI concept to be rewritten.
- Data exchange can evolve beyond conventional API-based architectures while retaining its core purpose: enabling governed movement and use of information.
This is what makes the proposed approach technology-agnostic. Therefore, the argument is not to allow today’s technology to determine tomorrow’s architectural boundaries.
From Technology Stacks to Sustainable Systems
Another observation is that DPI is sometimes discussed as though governments can acquire a defined stack of technologies and thereby acquire digital transformation. In practice, foundational infrastructure only creates possibilities. Its value depends on how registries are governed, how mediators are trusted, how data is exchanged, how institutions coordinate, how standards are set, and how systems integrate into wider service ecosystems.
This is particularly important for countries with constrained fiscal and institutional capacity. In many low- and middle-income countries with emerging digital ecosystems, the challenge is not simply deciding which DPI components to procure. It is determining what to build first, what to share, what to reuse, what dependencies exist, and what can realistically be sustained over time.
A functional approach helps make those decisions more explicit. It also helps countries avoid treating a globally recognised DPI architecture as an idealised stack that must be replicated in full. Instead, governments can assess their existing foundations, identify functional gaps and sequence investments according to their own institutional structures, development priorities and fiscal realities. This is not a rejection of global approaches. It makes them more adaptable to the country context.
Governance at the Heart of the Next Generation of DPI
Furthermore, as DPI evolves, the technical architecture cannot be separated from governance. Registries require clear mandates, custodianship and data governance. Mediators require trust frameworks, accountability, cybersecurity, auditability and rules governing how transactions and decisions are made. Data exchange requires standards, metadata, interoperability rules, privacy protections and controls over access and use.
This means that future-ready DPI is not simply an infrastructure agenda. It is also a governance agenda. Distinguishing BBs as functional capabilities from the outcome capabilities they enable can help governments make better decisions about policy, procurement, budgeting, and institutional responsibility. It can also reduce the risk that a particular vendor, technology or product becomes mistaken for the underlying infrastructure itself. That matters for digital sovereignty, as open, modular and standards-based architectures give governments greater ability to evolve their systems without becoming unnecessarily dependent on a particular technology stack or provider.
The first generation of DPI demonstrated the transformative potential of shared digital foundations. The next generation can make those foundations more durable. This does not require abandoning the concepts that have brought the field this far. Digital identity, digital payments and data exchange will remain critically important. Their importance is precisely why it is useful to understand the deeper functions that enable them.
The opportunity now is to move from defining DPI primarily through the outcomes that have become most visible, toward defining it through the functional capabilities that make those outcomes possible. That shift may appear subtle, but its implications are not. It changes how governments think about architecture, sequence investments, assess readiness, and approach procurement and standards. It also creates greater space for emerging technologies to become part of digital ecosystems without repeatedly forcing governments to redraw the foundations underneath them.
The future of DPI should therefore not be determined by which technologies are dominant today. It should be determined by whether the infrastructure we build today can support technologies, services, and forms of interaction we cannot yet fully anticipate. The goal is not to build DPI for the next technology. It is to build DPI that can evolve with the next technology.
This article draws on the author’s independent framework, “Reframing Digital Public Infrastructure Building Blocks: A Future-Ready Model for Emerging Technologies, Governance, and Development.” The full paper provides the detailed conceptual architecture underlying the approach.
*Leonora Smart-Abbey is an international advisor and strategist on public sector transformation, working at the intersection of governance, digital government, Digital Public Infrastructure (DPI), artificial intelligence, and institutional reform. She advises governments and development organisations on digital transformation and designing more trusted, resilient, and effective public institutions.
Her work combines strategic policy advisory, digital transformation, institutional design, and programme leadership to translate complex technological and governance challenges into practical, sustainable reforms. Her experience spans Africa, Europe, the Middle East, and the Pacific, with work across digital government, DPI, responsible AI, data governance, economic governance, and institutional reform. A defining feature of her work is her ability to connect domains often treated separately, helping governments move from fragmented digital initiatives and investments toward coherent, scalable, and sustainable transformation. (IDN-InDepthNews)