On August 17, 2026, the South African Revenue Service (SARS) published its Consultation Paper on VAT Modernisation: E-Invoicing, Interoperability Framework and E-Reporting. The paper sets out what SARS calls the Digital VAT Model, a system in which structured invoice data moves between suppliers, buyers, accredited service providers, and SARS in near real time. Stakeholders have until October 16, 2026, to submit comments.

The proposal follows the VAT Modernisation Discussion Paper that SARS released in 2023 and forms part of SARS Modernisation 3.0, the authority's transformation program for 2025/26 to 2029/30. Its stated end point is VAT auto-assessment, in which SARS uses trusted transaction data to prepare a VAT liability that the taxpayer then reviews, confirms, or amends.

The architecture is what makes the South African proposal stand out on the continent. Several African tax authorities require invoices to pass through a central government platform before they reach the customer. SARS is instead proposing a decentralized network in which accredited private access points validate invoices, with SARS connected to the network as the fifth corner. In this article, we explain how the model would work, when it would apply and which questions remain open for businesses, including foreign suppliers of electronic services.

Why SARS Wants to Move Beyond Post-Audit VAT

South Africa administers VAT as a self-assessed tax using the invoice-credit method. Registration, filing, and payment have already been digitized, but the consultation paper acknowledges that the system still depends heavily on unstructured invoice data, fragmented systems, and manual administration.

SARS currently verifies VAT returns through a post-audit model. It sees transactional activity only after a return has been filed, which the paper links to higher compliance costs, delayed certainty and refunds for taxpayers, and a reduced ability to tackle fraud and the VAT gap.

The Digital VAT Model is designed to reverse that sequence. Transaction data would reach SARS as invoices are exchanged rather than during an audit months later. SARS would combine that data with third-party information and use analytics to concentrate its interventions on exceptions, anomalies, and high-risk behavior.

Three Pillars: E-Invoicing, Interoperability, and E-Reporting

The model rests on three components that together form what SARS describes as a Decentralised Continuous Transaction Control and Exchange (DCTCE) model.

The first is the e-invoice itself. SARS defines it as a structured, machine-readable tax invoice using a prescribed data model, and the paper states plainly that a PDF, a scanned image, or an emailed document does not qualify. E-invoices would need to follow a recognized standard, and the paper names EN16931 CIUS, the UN/CEFACT Cross-Industry Invoice, and Peppol PINT BIS as examples. The detailed requirements would be set out in regulations, and SARS says the technical specifications will address zero rating, deemed supplies, and apportionment. The same rules would apply to electronic debit and credit notes.

Transactions that do not ordinarily require an invoice, such as deemed supplies and certain sector-specific supplies and deductions, fall outside the e-invoicing requirement as currently proposed. SARS says it will consider them during consultation and detailed solution design.

The second pillar is the Interoperability Framework (IF), a decentralized exchange network operated by service providers, which the paper compares to the providers that offer internet or email access. SARS identifies three conditions for the framework to work: systems must interoperate regardless of each taxpayer's IT environment, communication must be limited to authenticated participants, and all parties must follow common semantic and technical standards.

The third pillar is e-reporting through continuous transaction controls. VAT transaction data, including invoice data, would be submitted to SARS just before, during, or shortly after the supplier and buyer exchange it. Because SARS would have its own access point on the network, reporting would travel through the same framework as the invoices.

How the Five-Corner Model Works

The paper assigns a defined role to each of the five participants.

Corner One is the supplier. It must issue the e-invoice from accounting software capable of meeting the prescribed standard, appoint an accredited service provider from a list published by a Network Authority, and send each invoice to that provider in near real time.

Corner Two is the supplier's access point. It validates the invoice against the prescribed technical standards and VAT rules. A valid invoice is cleared and passed on to the buyer's access point and to SARS, while an invalid one is rejected and returned to the supplier for correction. SARS presents this step as the mechanism that stops invalid tax invoices at source.

Corner Three is the buyer's access point, which validates the invoice again on the buyer's behalf and delivers it to Corner Four, the buyer. The buyer's system processes the invoice automatically and responds with its VAT treatment of the purchase, indicating whether input tax is claimed in full, in part, or not at all. That response travels back through Corner Three to SARS. The paper calls this duplex clearance because both the supplier's and the buyer's access points report the same cleared invoice.

Corner Five is the access point appointed on behalf of SARS. It must connect to every service provider in the network and receive the transaction data of both suppliers and buyers for risk management and processing.

A Network Authority would accredit the access points and set and enforce the network's technical and legal rules. Each access point would contract with the Network Authority under a service level agreement.

SARS gives several reasons for choosing a decentralized design over a central clearance platform. With many providers in the network, the failure of one does not halt invoicing. SARS can choose to receive only a tax-relevant subset of each invoice after the service providers have checked data quality. SARS also needs to integrate only with the service providers rather than with every vendor, leaving the providers to handle most of the integration and support work with their clients.

From Pre-Filled Returns to Auto-Assessment

The data collected at Corner Five would first be used to pre-fill VAT returns and later to move toward automatic VAT assessment. Taxpayers would still have to confirm or edit the pre-filled figures, which SARS says preserves the self-assessment principle.

For finance and tax teams, the paper anticipates a change like the work. Instead of assembling data at month-end to prepare a return, teams would oversee automated invoice validation and resolve the exceptions the system flags. Internal controls would move earlier in the process, and the role of finance staff would shift toward data quality and exception handling.

Implementation Timeline

SARS proposes a five-phase journey. The durations are the paper's own estimates, and the paper describes the whole framework as subject to change following consultation and the relevant legislative and policy processes.

Phase 1, preparation, would begin in 2026/27 and run for about 12 months. It covers research, stakeholder consultation with industry bodies, service providers, and vendors, an assessment of readiness by segment, and the publication of draft VAT regulations. 

Phase 2, solution development, would take about 12 months in 2027/28 and would define the standards, specifications, and operating models for connectivity and data exchange, with the VAT regulations promulgated during this period.

Phase 3 would test the solution with voluntary participants in a controlled environment for about six months in 2028/29. 

Phase 4 would then run a pilot with voluntary participants from priority segments in a production-like environment for about six months in 2029/30. 

Phase 5, phased implementation, is expected to start during calendar year 2030 and extend over roughly 36 months, guided by mandates and turnover thresholds. Within this phase, large taxpayers and businesses would go first because their systems are more advanced. 

Government procurement would follow, possibly alongside the large-business segment, with public bodies required to accept e-invoices from suppliers. Micro, small, and medium enterprises would come next, with adoption spread over several years, and B2C transactions would come last. SARS says incentives may be considered to encourage consumers to accept and validate invoices.

The timeline is easy to misread. The paper does not set a date for mandatory e-invoicing. It says large taxpayers are expected to adopt the system voluntarily at first, with mandatory adoption introduced later, and that the order of segments may change depending on ease of adoption, compliance risk, and VAT gap indicators. SARS also expects legacy and new systems to run in parallel across segments until the transition is complete.

How South Africa's Approach Compares With Other African Mandates

South Africa's design differs from the clearance systems already running elsewhere in Africa. Ghana's E-VAT system requires each invoice to be sent to the Ghana Revenue Authority's Virtual Sales Data Controller for approval before it is issued to the customer. Ghana also requires non-resident digital businesses to issue e-invoices through their local clearance systems, which goes further than a registration obligation alone.

Kenya's eTIMS routes invoice data directly to the Kenya Revenue Authority. KRA has also added buyer-initiated invoicing, which allows a registered buyer to generate an invoice on behalf of a seller that cannot issue one itself, such as a small-scale farmer or an informal trader. That feature addresses the same problem South Africa's MSME phase will face: sellers who are part of formal supply chains but lack invoicing systems.

The closest reference point for South Africa is France, and the consultation paper notes that France plans a five-corner decentralized model for B2B e-invoicing and reporting. The paper's review of international lessons covers Mexico, Brazil, Chile, Italy, the EU's VAT in the Digital Age initiative, India, and China, but it does not discuss any African administration. That is a notable gap, given that several African tax authorities now have years of operating experience with invoice clearance, including in economies with large informal sectors.

The choice between the two approaches involves a real trade-off. A central platform gives the tax authority direct control over validation but concentrates operational risk in a single system. A decentralized network spreads the integration workload across private providers, but it only works if there is a sufficiently deep market of accredited access points that smaller businesses can afford.

Open Questions for Foreign Suppliers

The paper does not specifically address non-resident suppliers of electronic services that are registered for VAT in South Africa. Since April 1, 2025, foreign suppliers that supply electronic services exclusively to South African VAT-registered vendors fall outside the definition of electronic services. A foreign supplier that also sells to consumers, however, must still account for VAT on all its supplies regardless of the customer's registration status. Such suppliers would therefore issue tax invoices to South African businesses as well as consumers, and could in principle fall within the B2B phase. It is not yet clear whether they would be required to contract with a South African accredited access point or how their invoicing systems would connect to the network.

How to Respond to the Consultation

SARS is accepting written feedback through an online survey linked in the consultation paper until October 16, 2026. It specifically asks for input on the design of the Digital VAT Model, the phased implementation pathway, the readiness of taxpayers, software providers, and intermediaries, the costs, risks, and benefits of adoption, and the governance, standards, and safeguards needed for a trusted national system.

After the survey closes, SARS plans to consolidate submissions and hold structured working groups with associations, software vendors, government bodies, and taxpayers before adopting a final policy. Businesses that trade in South Africa, including foreign suppliers of electronic services and the platforms that serve them, should consider raising the treatment of non-residents, service-provider liability, data protection, and support for smaller vendors while the design is still open.

Conclusion

SARS describes the Digital VAT Model as a re-engineering of the VAT ecosystem rather than a technology upgrade, and the proposal supports that description. It would replace retrospective audits with continuous, network-based validation, assign defined roles to accredited private access points, and lay the groundwork for pre-filled returns and eventually auto-assessment.

Mandatory e-invoicing is still years away, and the paper leaves key questions unanswered, including who will run the network, which standard will apply, and how foreign suppliers and platforms will connect. Those questions will be decided over the coming months, which makes the current consultation the point at which businesses have the most influence over the outcome.