Nearpays wins UN AI for Good Innovation Factory

TL;DR

• Nigeria’s Nearpays won the AI for Good Innovation Factory 2026 Grand Finale in Geneva on 9 July, the first African startup to take the competition’s global title.¹²
• The International Telecommunication Union (ITU)-run programme awarded US$20,000; will.i.am separately matched that amount for each of the three finalists, bringing Nearpays’ total award funding to US$40,000.¹³
• Nearpays’ proposition combines NFC smartphone-based SoftPOS payments with an AI-led tax/compliance layer aimed at small merchants. That is a practical configuration for markets where terminal cost, connectivity and bookkeeping are genuine constraints.
• Do not confuse a respected pitch competition with deployment proof. Public reporting has not established Nearpays’ active-merchant count, payment volume, fraud performance, tax-authority integrations, or credit-access outcomes.

The award is real

Nearpays has won the UN AI for Good Innovation Factory’s 2026 Grand Finale, held during the AI for Good Global Summit in Geneva. The Innovation Factory is the ITU’s startup pitching and acceleration programme; it put Nearpays through its regional route in Johannesburg and then the global final.¹²

The company’s pitch centred on a SoftPOS system: compatible Android phones with NFC can accept contactless card payments without a dedicated terminal. Nearpays also describes an AI tax engine that uses transaction data to automate filing, optimise liabilities and flag potential over-taxation.¹⁴

That combination is why this is an AI story rather than merely a funding story. The claimed value is not “AI makes payments faster.” It is a closed operating loop: a merchant accepts a digital payment, the transaction becomes a business record, and that record is used for compliance and operational insight. If the loop works reliably, it can reduce the administrative cost of becoming legible to formal finance.

The quiet technical insight: data exhaust can become financial infrastructure

A cheap payment terminal is useful. A merchant’s reliable transaction history can be more valuable.

Many small businesses operate with little formal accounting history. The practical promise behind Nearpays is that payment acceptance produces machine-readable records, while a tax/compliance layer helps make those records usable. Over time, that could improve a merchant’s ability to document turnover and interact with lenders, payment partners or public authorities.

That is the non-obvious connection: the product is not only a POS replacement; it is a data-creation layer for businesses that have historically generated little usable financial data. The outcome, however, depends on factors outside the model: customer payment habits, acquiring-bank relationships, card acceptance, connectivity, tax rules, data governance and the willingness of lenders to use this data.

What the win means—and what it does not

The ITU’s jury selected a deployment-oriented AI product from a field that the organiser says moved from regional and thematic rounds to a Geneva finale.¹ That is a meaningful signal for African founders building applied systems around local infrastructure constraints rather than chasing a generic chatbot narrative.

But this is where award coverage usually becomes too generous. SoftPOS is an established payments category, not a new model breakthrough. Its hard part is execution: device security, card-scheme compliance, acquiring relationships, fraud controls, offline behaviour, reconciliation, merchant support and distribution. The AI tax layer adds another difficult dependency: correct tax treatment requires current rules, auditable logic, human review paths and usable links to relevant tax systems.

The award validates the problem framing and gives Nearpays visibility. It does not independently validate the accuracy of its automation, the security of its stack, the economics of acquiring merchants, or the claimed path from digital records to improved credit access.

Stakeholders: who gains if this works

Hype check: a prize is not a scale metric

The strong version of this story is that a Nigerian team built a credible product around African merchants’ actual constraints and won a competitive global platform. That is worth recognising.

The inflated version is that an award has “solved financial inclusion.” It has not. Nearpays has not publicly supplied the operational metrics needed to make that claim: active merchants, transaction volume, chargeback and fraud rates, uptime, average settlement time, tax-filing accuracy, or independently measured improvements in credit access. Until those appear, the correct label is promising applied fintech, not proven financial infrastructure.

Recommendations

For small-business owners considering smartphone payment acceptance: compare the full operating cost, not the hardware price. Ask the provider for merchant fees, settlement schedule, refund and dispute process, device requirements, offline behaviour, support hours and the precise data it retains or shares.

For financial institutions and payment partners: require evidence before treating transaction data as a credit signal. Test missing-data behaviour, fraud false positives, model drift, consent capture, appeal paths and the accuracy of tax recommendations. “AI-powered” is not a control framework.

For policymakers and development-finance actors: measure outcomes at the merchant level: new payment acceptance, retained use after 90 and 180 days, transaction reliability, effective fee burden, documentation quality and access to appropriately priced credit. A pitch-stage award is a useful lead, not an impact evaluation.

Uncertainty ledger

• Public sources confirm the award, the prize structure and Nearpays’ stated product proposition. They do not independently disclose merchant scale, payment volume or unit economics.
• Nearpays and AI for Good describe AI functions in payments, operations, compliance and tax automation; detailed technical documentation, model methods and independent security assessments were not located in this review.
• The claim that digital payment records can improve access to credit is plausible but outcome-dependent. This analysis would become materially stronger with audited adoption data and a named lending or tax-authority integration.

Bottom Line

Nearpays won the UN AI for Good Innovation Factory because it addresses a real, specific bottleneck: the cost and friction of bringing small merchants into digital payments and record-keeping. The win matters as evidence that African applied-AI products can set the agenda on a global stage. It does not yet prove the company has solved the harder problem—secure, trusted, high-volume deployment that improves merchants’ financial lives.

Sources