Agents of Mobile Autonomy: Leveraging Agentic AI in East Africa’s Mobile Payment Services

In the fast-evolving landscape of Artificial Intelligence (AI), agentic AI stands as the next frontier, enabling systems to autonomously pursue user goals beyond simple query responses. Unlike traditional assistants such as scripted chatbots, agentic AI drives proactive decisions, tool usage, and iterative planning, emulating human agency in dynamic settings. For East Africa, where mobile money has sparked financial inclusion, this innovation could be revolutionary, offering transformative potential for the region’s mobile payments while demanding strategic handling of infrastructure and regulatory challenges.

Mobile Money Landscape in East Africa

East Africa spearheads global mobile money adoption, with its high penetration filling the gap left by traditional bankingand driving regional financial inclusion. By 2022, the number of mobile money users in Eastern Africa was the highest on the continent, with the region reporting 115 million active accounts out of the 390 million registered.

Kenya exemplifies this as its mobile money giant, Safaricom's M-Pesa, serves over 60 million monthly active customerswith transactions, remittances, and micro-lending amid weak banking networks, not just in Kenya but in other countries such as Tanzania, Ethiopia, and the Democratic Republic of Congo(DRC). MTN Mobile Money (MoMo) is another mobile money provider that offers integrated services like peer-to-peer transfers, merchant payments, banking services, and digital commerce to millions across 16 countries, including Uganda and Rwanda. Tanzania’s Tigo Pesa interconnects regionall,y whereby its users can transact with users of Safaricom's M-Pesa in Kenya, MTN Mobile Money in Uganda and Rwanda, and Airtel Money in Rwanda, hence enhancing accessibility.

Such interoperability pacts now enable seamless transfers across Kenya, Uganda, Tanzania, and Rwanda, slashing costs for remittances and trade. Additionally, once fully operational, the Rwanda-Tanzania pilot by the East African Payment System (EAPS) will link national payment systems to enable instant, low-cost transfers between financial institutions in participating countries. However, Agentic AI could elevate these initiatives by automating fraud detection, tailoring financial advice, and streamlining cross-border payments for greater efficiency.

Understanding Agentic AI

Agentic AI comprises autonomous systems that reason, plan, and execute actions with minimal human oversight, transforming passive tools into proactive partners. These agents define goals, wield external tools for tasks like data retrieval or automation, and refine strategies through real-time feedback, adapting to dynamic environments. This marks a leap from reactive chatbots limited to query responses to decision-making entities that orchestrate multi-step processes independently, fostering efficiency in enterprise and consumer applications.

Foundation models have advanced with extended time horizons. Anthropic's Claude Sonnet 4.5, for instance, autonomously completes over 30 human hours of work in coding and agentic workflows like office automation.

Multi-agent collaboration systems now power complex workflows via protocols like Agent-to-Agent (A2A), enabling seamless coordination, negotiation, and dynamic task delegation across agents.

Finally, integrating large language models (LLMs) with reinforcement learning and continual learning through standards like Model Context Protocol (MCP) equips agents with persistent memory and adaptive reasoning, supporting multiday tasks and reducing intervention needs.

Global Impact on E-Commerce

Globally, Agentic AI has already had a transformative impact on e-Commerce by enabling autonomous agents to handle discovery, negotiation, and transactions, shifting from human-led browsing to AI-orchestrated experiences. For instance, a survey by Adobe Analytics found that U.S. retail site traffic from Generative AI browsers surged 4,700% year-over-year in July 2025. Notably, findings also indicated that use of AI tools leads to more informed customers, with 73% citing Generative AI as their primary source of product research and 85% of respondents noting it improved their shopping experience.

This surge signals a paradigm where AI agents act as proactive shoppers, as demonstrated by Alibaba's AI Mode, which interprets natural language queries to compare suppliers on pricing and logistics. Additionally, Google's agentic checkout tracks prices and completes buys autonomously using Google Pay, enhancing holiday efficiency for merchants like Wayfair.

These Agentic AI shifts promise efficiency, which, when similarly integrated into East Africa’s Mobile Money landscape, could reshape the industry by creating a surge in efficiency through a technological redefinition of service delivery.

AI Application in the Regional Mobile Money Landscape

Opportunities

Operational efficiency and cost reduction.

Agentic AI integration into mobile money services could enable 24/7 natural-language handling of inquiries such as balance check, PIN resets, process micro-transactions, reconcile settlements, and route cross-border payments without oversight. This is already being implemented globally by autonomous shopping assistants like OpenAI's Operator and Perplexity's Buy with Pro, which have been able to automate end-to-end purchases, integrate research, recommendations, and checkouts via partnerships with Shopify and PayPal. This could minimise call centre workloads and lead to faster transaction cycles, hence cutting delays and lowering operational costs.

Enhanced customer experiences via personalised services.

Personalised recommendations leverage browsing history and real-time behaviour for hyper-targeted suggestions. By analysing individual mobile data and spending patterns, Agentic AI could curate savings products, dynamic micro-lending via utility and usage metrics, and customise insurance as well as investments, hence boosting uptake among underserved users. Real-time analytics could also forecast trends, automate reporting, and predict behaviours for agile service evolution.

Furthermore, personalised services could accelerate financial inclusion in East Africa through conversational banking with multi-language support, such as the incorporation of Swahili, the inclusion of linguistic dialects, and voice-activated transactions, which aid low-literacy groups, while WhatsApp-integrated AI guides could enable seamless payments. Underserved groups could also benefit from AI credit assessments for the unbanked via mobile data, automated micro-savings, and predicted crop cycles for agricultural payouts. Merchant Services AI could also enable easy payment acceptance, inventory forecasting, and dynamic pricing for small businesses, fostering growth.

Fraud Detection and Security Enhancement

Real-Time Threat Monitoring AI agents could scan patterns to flag anomalies in milliseconds, adapting to tactics with biometric verification for robust security in high-volume mobile flows. Risk Management Mobile history could also fuel alternative credit scoring, automating compliance and predicting defaults to minimise losses.

Possible Implementation Barriers

Digital Infrastructure Deficits

East African countries such as Kenya, Tanzania, Uganda, DRC, and Ethiopia are among those with the widest digital connectivity gaps on the continent. These statistics reinforce the findings that Africa has the largest mobile connectivity gap globally. Additionally, computing infrastructure is also lacking, with Africa claiming just 1% of global AI compute capacity and less than 2% data centres, limiting model training and hardware access. Integration Complexity Legacy incompatibilities, platform silos, and absent API standards also hinder mobile money-AI fusion.  

This could be addressed by prioritising offline-capable AI, like Uganda's health supply chain framework, which addresses low-connectivity challenges in rural areas. Public-private partnerships (PPPs) are also essential for AI infrastructure, similar to the World Bank's Cotonou Declaration, which mobilises blended finance for digital investments in Western and Central Africa. Additionally, phased implementation via pilots, such as OECD-supported AI enhancements for mobile money credit scoring in East Africa, ensures scalable rollout.

Regulatory and Governance Challenges

There is a fragmented regulatory environment in East Africa due to a lack of formal AI laws, with only Kenya and Rwanda having national AI strategies as of mid-2025. Even more, the AU's Continental AI Strategy, adopted in 2024, enters Phase I implementation in 2025 but awaits full ratification. Enforcement gaps persist despite Data Privacy and Protection laws, while sovereignty and cross-border flow concerns undermine AI trust. This has led to adoption barriers such as fluctuating consumer confidence with fears of AI bias in lending, cultural aversion to autonomy, and opacity concerns, thus widening digital divides. To address this, regional coordination that harmonises regulations for ethical AI development is necessary.

Human Capital and Financial Constraints

The significant technical skills gap is evident, as demonstrated by the existence of AI talent clusters in Nigeria, South Africa, Kenya, Egypt, and Morocco, which comprise over half of Africa's developers, thus indicating training and literacy lag elsewhere in the region. The costs of bridging this gap, as well as AI infrastructure costs, exacerbate barriers, considering a plunge in Africa’s fintech funding that is from 45% in 2024 to $857 million, with 70% sourced externally. This can be addressed by regional investment in AI literacy through UNESCO's national competency frameworks and local-language programs to build capacity across the continent.

Conclusion

AI is forecasted to add $1.2 trillion to Africa's economy by 2030, per Microsoft estimates. While Agentic AI could unlock transformative potential for East African mobile money, enhancing inclusion and efficiency, success hinges on aligning innovation with robust infrastructure and mature regulations to address hurdles like spotty infrastructure, privacy risks, and regulatory flux. Through strategic, collaborative efforts, East Africa can leapfrog legacy banking, spurring economic growth. Stakeholders must prioritise responsible AI to equitably harness this frontier.