The commercial rollout of artificial intelligence has entered a distinct new phase: the hunt for hard ROI. As businesses recognize the hidden costs of AI deployments, CFOs are mandating stricter cost-to-value ratios before signing off. Today's briefing unpacks that shift, alongside a structural warning about the 'Know Your Agent' compliance gap in financial services, and a detailed look at how Weaver Fintech is using 'Buy Now, Pay Later' to acquire South African credit customers at single-digit costs.
Following recent data highlighting the hidden costs of 'botsitting' and low-quality outputs, a new Times of India analysis confirms a broader market shift: enterprises are moving from rapid, experimental AI deployment to a disciplined, cost-conscious phase focused on measurable business returns. Major operators like Uber, Meta, and Amazon are reportedly implementing cost controls and usage caps as the marginal returns from advanced models diminish. Meanwhile, Indian organizations are planning a 45% increase in AI investment, but with a strict new focus on ROI.
Why it matters
This marks a crucial maturation point for enterprise AI. The era of 'deploy at all costs' is ending, replaced by CFO-level scrutiny on unit economics. For operators, this means the commercial conversation is no longer about an AI model's capability, but its cost-to-value ratio. This will force AI vendors to compete on efficiency and demonstrable ROI, likely leading to more vertical-specific, cost-optimized models and a clear preference for solutions that produce tangible results over speculative gains.
Dilip Asbe, the CEO of India's National Payments Corporation (NPCI), stated on Monday that AI will be the central pillar for the next phase of growth for the Unified Payments Interface (UPI). The goal is to surpass one billion daily transactions by leveraging AI for user growth, enhanced fraud prevention, and credit distribution. Asbe specifically mentioned the potential for Indian financial institutions to build their own small language models (SLMs) on rich domestic data and the use of AI for dispute resolution.
Why it matters
This is a roadmap for how to scale a public payments utility. The head of one of the world's most successful payment infrastructures is explicitly stating that the path from 500 million to 1 billion daily transactions runs through AI. For operators in other emerging markets, this provides a blueprint for using AI not just for chatbots, but for core operational challenges like fraud at scale, onboarding the next wave of users with multilingual voice interfaces, and even powering agentic commerce.
A comprehensive guide published on Monday provides a snapshot of the leading AI agent platforms for business automation in 2026, a market projected to hit $10.9 billion. The analysis compares top players like FwdSlash, Salesforce Agentforce, and Microsoft Copilot, detailing their features, pricing models, and specific use cases in sales, marketing, and data integration. The focus is on how enterprises are using these agents to automate complex workflows and enhance customer experiences.
Why it matters
This provides a practical map of the vendor landscape for AI agents, moving beyond theoretical capabilities to what is actually available and being deployed. For an operator planning to leverage AI, this is a useful starting point for vendor selection and understanding the different strategic approaches in the market—from all-in-one platforms like Salesforce to more modular tools. It highlights the rapid maturation from 'what is an AI agent?' to 'which AI agent should we use?'.
South African fintechs, notably Weaver Fintech (the group behind PayJustNow and Finchoice), are successfully challenging traditional credit providers by using Buy Now, Pay Later (BNPL) as a low-cost customer acquisition engine. A new analysis details how they acquire customers for as little as R9 through BNPL offerings, then leverage the transaction data to profile users and cross-sell higher-margin lending and insurance products. This digital-first risk assessment and ecosystem approach allows them to bypass the constraints of incumbent banks.
Why it matters
This is a powerful case study in operator strategy, demonstrating a playbook that will be highly relevant to the South African market. Instead of competing head-on with banks on their terms, these fintechs are using a Trojan horse. The low friction of BNPL builds a valuable customer base and proprietary dataset that incumbents, with their legacy systems and regulatory burdens, cannot easily replicate. This is a clear example of a distribution advantage creating a new credit ecosystem.
Following its recent upgrade to a national license by the Central Bank of Nigeria, fintech giant OPay has opened a new regional office in Kaduna State, signaling a significant push to deepen its presence in northern Nigeria. The new hub, inaugurated on June 10, will coordinate customer support, merchant services, and local partnerships. The move is part of a broader nationwide expansion strategy focused on driving financial inclusion beyond the primary commercial centers.
Why it matters
This is a classic operator move demonstrating that even in the digital-first world of fintech, scaling in a market like Nigeria requires a physical presence. By establishing a regional hub, OPay is investing in local relationships, merchant trust, and last-mile support—key ingredients for winning in underserved markets. For competitors, it's a reminder that a slick app isn't enough; building an on-the-ground network is crucial for long-term market penetration and defensibility in Africa.
Nigeria's President Bola Ahmed Tinubu has approved the free corporate registration of 250,000 Micro, Small, and Medium Enterprises (MSMEs). The initiative, run through the Corporate Affairs Commission (CAC), aims to integrate a huge number of informal businesses into the formal economy by removing the financial barrier to registration, which represents about N3 billion in foregone fees.
Why it matters
This is a massive expansion of the addressable market for any fintech or B2B service provider in Nigeria. Bringing a quarter-million businesses into the formal economy overnight creates a surge in demand for bank accounts, payment processing, accounting software, and credit. For operators focused on merchant services, this government-led formalization push represents a significant, non-organic growth opportunity that dramatically widens the top of the sales funnel.
Expanding on the 'intent gap' and early 'Know Your Agent' signaling from the FCA that we've been tracking, a new analysis argues that traditional compliance architecture is now effectively obsolete. Current Know Your Customer (KYC) and identity management (IAM) frameworks were built for human actors and are ill-equipped to govern non-deterministic, self-spawning AI agents. This creates a massive structural gap, exposing firms to significant governance failures and regulatory risk as agents increasingly initiate financial transactions.
Why it matters
This identifies a critical, second-order problem created by the rush to deploy agentic AI. While the industry focuses on what agents *can* do, the regulatory and compliance infrastructure for what they *are* doing is missing. This isn't just a theoretical risk; it represents a major new category of RegTech that will need to be built. For any operator deploying AI in a regulated space like payments, solving the 'KYA' problem will be a prerequisite for operating at scale, creating both a compliance burden and a significant product opportunity.
Business owners in South Africa's township tourism sector are reporting significant booking cancellations ahead of planned demonstrations on June 30. The protests, which are reportedly focused on immigration, are creating safety concerns among international and regional tourists, threatening the fragile post-pandemic recovery of this vital part of the hospitality economy.
Why it matters
This is a direct hit to discretionary spending and footfall in a key segment of the local economy. It's a reminder of how quickly social and political instability can translate into economic pain for small businesses. For retailers and hospitality operators, it underscores the vulnerability of consumer behavior to the broader socio-political climate, impacting revenue and planning far beyond the immediate areas of protest.
To mark the 35th anniversary of Sonic the Hedgehog, Sega has collaborated with Jersey Jack Pinball on a new pinball table. Designed by legendary pinball designer Steve Ritchie, the machine features an HD LCD display and immersive audio. The collector's item is priced at $9,999.
Why it matters
This release taps into the powerful commercial force of retro gaming nostalgia, translating a classic digital IP into a high-end physical product. It's a great example of how enduring brands from the 8-bit and 16-bit eras continue to find new, premium monetization channels by catering to a dedicated and affluent fanbase.
Building on the stablecoin-first settlement architectures we've seen adopted by Flutterwave and Daya, a new report highlights that African Payment Service Providers (PSPs) are systematically bypassing the traditional SWIFT network. Firms are increasingly using institutional OTC desks to achieve stablecoin settlement in hours instead of days, gaining cost predictability and reducing foreign exchange exposure.
Why it matters
This is a significant infrastructure leapfrog moment for pan-African commerce. Rather than waiting for legacy systems to improve, operators are actively building and using a parallel financial system for B2B settlement. It's a pragmatic response to the real-world friction of cross-border trade in Africa. This trend, enabled by a maturing regulatory view of crypto-fiat settlement in key markets, could fundamentally reshape the economics of intra-African and international trade for businesses on the continent.
Enterprise AI Enters a Cost-Conscious Era Across multiple reports, a clear trend is emerging: enterprises are moving beyond the initial hype cycle of AI deployment and are now intensely focused on cost control and measurable ROI. Companies like Uber and Meta are implementing usage caps, and investment strategies are being re-evaluated to prioritize demonstrable business value over purely technical capability (c_36).
The 'Know Your Agent' Compliance Gap As autonomous AI agents begin to execute financial transactions at scale, a new structural problem is coming into focus: the inadequacy of existing KYC/IAM frameworks. A new analysis argues for a 'Know Your Agent' (KYA) compliance layer to govern non-human actors, creating a significant new challenge and opportunity in the regulatory tech space (c_37).
BNPL as a Low-Cost Customer Acquisition Funnel A case study from South Africa shows how fintechs are strategically using Buy Now, Pay Later not just as a product, but as a highly efficient customer acquisition channel. By acquiring users at a very low cost (as little as R9), firms like Weaver Fintech are building detailed profiles to cross-sell higher-margin lending and insurance products, outmaneuvering traditional credit providers (c_11).
The 'SaaSpocalypse' Accelerates The thesis that AI will fundamentally disrupt the traditional SaaS model continues to gain evidence. With AI agents making it easier to build software and enabling outcome-based pricing, the long-standing per-seat, 'access-to-tools' model is under pressure. This is forcing a reevaluation of SaaS valuations and go-to-market strategies (c_39).
Digital Expansion in Nigeria Requires Physical Presence Despite the digital-first nature of fintech, players like OPay are demonstrating that deep market penetration in Nigeria requires a physical footprint. OPay's new regional office in Kaduna highlights a strategy of building localized support and merchant relationships to drive financial inclusion and trust in underserved areas (c_61).
What to Expect
2026-06-30—UNCTAD report on digital tech for MSMEs in East Africa to be released.
2026-08-01—EU AI Act's new transparency rules for AI-generated content in retail take effect.
— The Merchant Desk
🎙 Listen as a podcast
Subscribe in your favorite podcast app to get each new briefing delivered automatically as audio.
Apple Podcasts
Library tab → ••• menu → Follow a Show by URL → paste