Digital Identity and the Next Phase of Financial Inclusion in South Africa
AO commentary on Facephi analysis of multisectoral fintech in South Africa
Facephi has published an analysis of collaborative fintech and financial inclusion in South Africa. AO looks at inclusion as an architecture problem: identity, payments, data, risk, channels and regulation have to connect before access becomes useful.
Based on an original publication by Facephi
Facephi has published an analysis of South Africa's fintech ecosystem and the role of multi-sector collaboration in financial inclusion. Where that piece maps the ecosystem, we want to look at the engineering consequence, because inclusion programmes in our markets rarely stall on ambition. They stall on integration.
A note on evidence: any market statistics and regulatory characterisations in the source article belong to Facephi and its own sources. We have not independently verified them, so we have not repeated them here as AO research. Where you need a number for a business case, go to the primary source and check its date and definition.
Financial inclusion is more than opening an account
An account that receives a grant and is emptied at an ATM the same day has not included anyone in the financial system. It has digitised a cash withdrawal.
Useful inclusion means access to payments that work for how people actually trade, a safe place to keep value, credit that is priced on evidence rather than assumption, insurance that pays out without a month of paperwork, and enough visibility for a household or a small business to plan. Each of those depends on a different part of the stack, which is why inclusion is an architecture question rather than a product launch.
Identity is the entry point
Everything downstream is gated by one thing: can the institution establish, remotely and proportionately, who this customer is?
Where the only credible route is a branch visit with original documents, geography does the excluding. Remote verification, combining document capture, liveness and biometric matching, changes the economics of reaching a customer who lives two taxi rides from the nearest branch. It has to be implemented within the institution's regulatory obligations, and it has to allow for assisted and lower-friction paths for customers with older devices or limited connectivity.
Alternative data can create new financial signals
Thin credit files are the second gate. A customer with a decade of reliable mobile-money activity, consistent merchant purchases, a paid-up energy account or a documented agricultural cycle is not high risk. They are unmeasured.
Conceptually, several signal families can help close that gap: mobile transaction and payment behaviour, merchant and till activity, agricultural production and input data, and utility or prepaid energy payment history.
We would be careful here. AO does not suggest that every available dataset should be used because it exists. The tests we would apply are consent that a customer genuinely understands, purpose limitation, proportionality, fairness and bias testing on the resulting decisions, explainability to the customer and the regulator, and compliance with POPIA and applicable sector rules.
Behavioural intelligence after the first transaction
Once a customer is verified and transacting, the institution stops guessing. Behavioural intelligence can help it understand income and expenditure rhythms, the goals a customer is saving towards, when a product no longer fits, and when risk is changing, and can inform engagement that is useful rather than generic.
This is also the point at which inclusion and fraud prevention converge, because the same behavioural view supports both understanding a customer and noticing when an account stops behaving like them. Our commentary on synthetic identities and mule accounts covers that side.
AO's connected model
We use the same sequence for inclusion that we use for connected banking generally:
- Identify & Onboard
- Protect & Monitor
- Understand & Engage
- Grow & Monetise Responsibly
Inclusion fits the model exactly. Identify and onboard is where reach is won or lost. Protect and monitor is what makes low-value, high-volume accounts safe to operate. Understand and engage is how a first product becomes a relationship. Grow and monetise responsibly is the discipline that keeps the last step from undoing the first three.
Data sovereignty and governance
Programmes of this kind attract scrutiny for good reason. The practical design inputs we raise early with clients are POPIA obligations and lawful basis, explicit and revocable customer consent, data minimisation in what is collected and retained, where data is processed and stored and who can reach it, security architecture for biometric and behavioural data, and governance and approved infrastructure for any AI in the decision path.
None of that is legal advice, and it is not a substitute for the institution's own compliance assessment. It is the set of questions that changes what gets built.
Where AO fits
AO is the integration and delivery layer. In inclusion programmes that usually means connecting identity providers, core banking and payment rails, transaction monitoring, data platforms and AI services, digital and USSD or assisted channels, and the professional services to run it afterwards.
Explore Connected Banking, Behavioural Banking, our Identity and eKYC capability or systems integration.
Frequently asked questions
- Why is digital identity important for financial inclusion?
- Regulated financial products require an institution to establish, to a proportionate standard, who its customer is. Where in-branch document checks are the only route, distance and cost exclude people. Remote identity verification, including document capture and biometric matching, makes it possible to onboard customers who cannot easily reach a branch, subject to applicable regulation such as South Africa's FICA obligations.
- How can remote onboarding expand access?
- It removes travel, queueing and paperwork from the first interaction, and it allows an institution to serve customers in areas where a physical footprint is uneconomic. It also shortens the gap between intent and activation, which is where most inclusion funnels lose people. It is not a universal answer: device access, connectivity, digital literacy and assisted channels still matter.
- What role can alternative data play?
- Where formal credit history is thin, other signals can help build a picture of financial behaviour, for example payment and mobile transaction patterns, merchant activity or utility and energy payment records. These signals can support responsible credit and product design, but they should be used with explicit consent, clear purpose limitation, fairness and bias testing, and within the requirements of POPIA and sector regulation.
- How should financial institutions govern biometric and behavioural data?
- Treat it as sensitive by default. In practice that means collecting the minimum needed, being explicit about purpose and retention, controlling where data is processed and stored, securing it in transit and at rest, restricting internal access, documenting model and vendor oversight, and giving customers a meaningful way to understand and exercise their rights. Specific obligations depend on the regulation that applies to the institution.
- Why does financial inclusion require multi-sector collaboration?
- Identity, distribution, data, trust and payment rails typically sit with different organisations. Banks hold regulated capability, mobile operators hold reach, fintechs hold product velocity, government holds foundational identity, and sector players such as agriculture, retail and energy hold behavioural data and customer relationships. Programmes that connect these assets tend to reach further than programmes that try to own all of them.
Source & attribution
This article contains AO commentary based on an original publication by Facephi. The underlying announcement and statements regarding Facephi, its technology and partnerships originate from Facephi.
The source analysis, including any market statistics and regulatory descriptions it contains, originates from Facephi. AO has not independently verified those figures and does not present them as AO research. Nothing here is legal or regulatory advice.
- Original source:
- Facephi
- Original publication:
- Multisectoral Fintech in South Africa: Collaborative Technologies for Financial Inclusion 2026
- Original author:
- Facephi Observatory
Sources and references
- Multisectoral Fintech in South Africa: Collaborative Technologies for Financial Inclusion 2026 — Facephi. Original analysis by the Facephi Observatory. Statistics and regulatory descriptions in that article are attributable to Facephi and its own sources.
Planning something like this?
Book a Discovery Call with the AO team and we will work through it with you.
Related from AO
Synthetic Identities and Mule Accounts: Why KYC Alone Is No Longer Enough
Facephi has published research on detecting mule accounts opened with synthetic identities. AO looks at what that means architecturally: identity verification and transaction monitoring can no longer be run as two unrelated controls.
ThitsaWorks and Tazama Partner to Advance Real-Time Fraud Protection
The ThitsaWorks and Tazama partnership highlights an important challenge for digital finance: expanding access and accelerating payments while strengthening the technology used to identify and respond to financial crime.
Tazama and WiredIn Partner Following Joint RSwitch Deployment in Rwanda
A strategic partnership between Tazama and WiredIn, following their first joint deployment with RSwitch in Rwanda, provides a practical example of how modern fraud and transaction-monitoring capabilities can be integrated into Africa's evolving digital payment ecosystems.