Samuel Adekunle on Why Data Sovereignty Will Shape Nigeria’s AI Future

Ugo Aliogo

As artificial intelligence becomes embedded in banking, healthcare, telecommunications and public services, questions about where data is stored and who controls the systems processing it, are becoming increasingly important.

Samuel Adekunle, Co-founder of BimpeAI, is contributing to this emerging field by developing enterprise AI infrastructure for African organisations. BimpeAI builds voice and multichannel AI agents that connect with existing business systems, support customer interactions and execute operational workflows across voice, messaging and digital channels.

Adekunle’s work focuses on moving artificial intelligence beyond demonstrations into production environments. His areas of interest include voice infrastructure, systems integration, real-time workflow orchestration, African-language AI and deployment models designed for organisations operating in regulated industries.

Through BimpeAI, he is working to address a persistent challenge in enterprise AI: enabling institutions to benefit from automation without surrendering control over sensitive customer and operational data.

Under his leadership, the company has [insert verified information about deployments, customers, markets, interactions processed, partnerships, revenue growth or measurable results].

In this interview, Adekunle discusses Nigeria’s journey from digitisation to artificial intelligence, the growing importance of data sovereignty and why locally relevant AI infrastructure could become strategically important to the country’s economy.

Nigerian businesses have moved from basic digitisation to cloud adoption over the past decade. How do you see that journey?

The first phase was about helping organisations move online. Banks, retailers, schools, hospitals and government agencies began replacing paper-based and manual processes with digital systems.

The next phase was cloud adoption. Organisations wanted greater speed, lower infrastructure costs and the ability to scale without purchasing new servers every time they introduced a product or expanded a service. Cloud technology helped many Nigerian companies innovate and reach customers more efficiently.

Artificial intelligence introduces another layer to that journey. The question is no longer only, “Where is our data stored?” Institutions must now ask, “Who is processing and reasoning over our data? Where is that processing taking place, and what happens to the information after an AI system interacts with it?”

This is why the conversation is moving from cloud adoption towards data sovereignty. AI has transformed data infrastructure from a storage question into an issue involving control, accountability and institutional trust.

Many people consider data residency to be primarily a regulatory concern. Why does it matter in practical terms?

Data residency refers to the physical or geographical location in which data is stored and managed.

For an individual user, this may sound abstract. For a bank, hospital, telecommunications company or public institution, however, data represents people’s identities, finances, medical records, transactions, addresses, complaints and behavioural information.

When this information is stored or processed outside the country, institutions must consider several questions. Which jurisdiction’s laws apply? Who may obtain access to the data? Can Nigerian regulators audit how it is being handled? What protections exist if there is a breach or operational dispute?

Data residency is therefore not merely a technical or compliance matter. It concerns an organisation’s ability to maintain control over critical information and retain the trust of the people it serves.

The Central Bank of Nigeria has directed financial institutions and payment participants to store and manage payment transaction data generated in Nigeria locally from January 2027. What does this signal?

It indicates that Nigeria is beginning to treat digital sovereignty as a strategic priority.

The financial sector is one of the most sensitive components of the economy. Payment data can reveal how individuals and businesses earn, spend and transfer money. It is understandable that regulators want stronger oversight over where such information is stored and managed.

The directive also has implications for AI companies. If payment, customer, KYC and transaction data require stronger local control, AI systems that interact with that information must be designed accordingly.

It is no longer possible to discuss the responsible deployment of AI in banking without considering where the associated data is stored, processed and logged. The two conversations have become inseparable.

How does artificial intelligence change the data-residency conversation?

Before the rapid adoption of AI, data residency was primarily concerned with the location of databases, servers and backups. Artificial intelligence makes the issue more complex because an AI system does more than store information.

It can read data, identify patterns, summarise records, produce recommendations, generate responses and, in some cases, initiate actions through connected systems.

For example, when a bank uses an AI agent for customer support, collections, onboarding or account-related services, the system may interact with sensitive customer information. The institution must understand where the AI is operating, where prompts are sent, whether interactions are logged and whether customer information leaves its controlled environment.

It must also determine whether a model provider can retain or reuse the data and whether the institution can reconstruct and audit the AI’s actions.

These are increasingly important questions for organisations moving from experimental AI projects to production deployments.

Global AI platforms are already highly capable. Why should Nigeria invest in local AI infrastructure?

Global AI platforms are powerful and valuable. The issue is not whether they should be used, but whether every platform or deployment model is appropriate for every Nigerian use case.

Using an AI tool to draft general marketing content presents a different level of risk from using AI to speak with a bank customer, process KYC information, examine financial behaviour or support a government service.

In regulated and high-trust environments, institutions require greater operational control. They need to understand where a model is hosted, what information it can access, what actions it is authorised to perform and whether its decisions can be audited under Nigerian law.

Nigeria therefore needs local infrastructure and deployment options that allow organisations to benefit from global advances in AI while retaining appropriate control over sensitive data and critical workflows.

That is where sovereign AI becomes important.

What do you mean by “sovereign AI”?

Sovereign AI is a country’s ability to develop, host, govern and deploy artificial intelligence in ways that reflect its laws, languages, infrastructure and national priorities.

For Nigeria, this includes systems that can understand Nigerian English, Pidgin and languages such as Yoruba, Igbo and Hausa, including the accents and code-switching patterns that shape everyday communication.

It also means developing AI infrastructure that can operate close to sensitive data, particularly in banking, healthcare, telecommunications, identity management and public services.

Sovereign AI does not mean isolating Nigeria from international technology or rejecting global models. It means creating the domestic capability required to use those technologies safely, adapt them to local realities and maintain control over strategically important systems.

What concerns are enterprises, particularly banks, raising about AI deployment?

The appetite for AI among Nigerian enterprises is significant. Banks and other large institutions understand that AI can reduce pressure on contact centres, improve customer engagement, support collections, reactivate dormant customers and make internal processes more efficient.

Their principal concern is governance.

They want to know whether customer data will leave their environment, whether an AI solution can be deployed on-premise or within an approved cloud, and whether it can connect with existing systems without unnecessarily exposing sensitive information.

They also want mechanisms for controlling what an AI agent is permitted to say or do. They need access to logs and a reliable record of the system’s interactions and actions.

These concerns should not be interpreted as resistance to innovation. They are the questions responsible organisations must answer before deploying AI at scale.

What categories of Nigerian data require particular protection when AI is involved?

Information connected to identity, finance, health, access to essential services or an individual’s rights should be handled with particular care.

This includes BVN- and NIN-linked information, KYC documentation, transaction histories, loan records, customer complaints, health information, biometric data, citizen records and confidential enterprise information.

The risk is not limited to whether data is securely stored. An AI system may infer sensitive information from a dataset, reveal it in an inappropriate response or take an action without the necessary authorisation.

AI governance must therefore go beyond traditional data protection. It should address access controls, model behaviour, human oversight, output monitoring, action permissions and accountability throughout the system’s lifecycle.

How does on-premise AI address some of these concerns?

On-premise AI allows a model or AI application to operate within an organisation’s environment or approved local infrastructure instead of routinely transmitting sensitive information to an external platform.

For banks and other regulated enterprises, this can provide greater control over data access, security policies, interaction logs and compliance requirements. It can also facilitate integration with internal systems while limiting unnecessary exposure of confidential information.

At BimpeAI, this is an important area of product development. We see an opportunity to build AI agents and voice technologies that operate closer to enterprise data, particularly in workflows where security, reliability and regulatory oversight are essential.

However, on-premise deployment is not automatically secure. It must still be supported by proper access controls, testing, monitoring, cybersecurity practices and clearly defined human accountability.

BimpeAI is developing voice AI for Nigerian and African applications. How does voice connect with sovereign AI?

Voice is central to the sovereign AI conversation because language itself is a form of infrastructure.

If AI systems cannot reliably understand Nigerian accents, Pidgin, local languages and code-switching, they may exclude a considerable part of the population or provide a poor standard of service.

Many globally developed speech systems were not primarily trained around the diversity of Nigerian speech. This can affect their performance when users switch between languages, speak in noisy environments or use expressions that are specific to a local context.

In banking and public services, these limitations can have serious operational consequences. A voice agent must understand what a customer is requesting, distinguish casual conversation from an instruction and respond appropriately.

BimpeAI’s work in this area is focused on making voice-based AI more relevant to Nigerian and African enterprise environments. This involves improving how systems handle local speech patterns, mixed-language conversations and the practical conditions in which customers communicate.

Building effective systems for these environments requires local knowledge, carefully governed datasets, continuous evaluation and product development that reflects how people actually speak.

What is your role in developing BimpeAI’s technology and commercial direction?

As Co-founder, my responsibility is to help translate advances in artificial intelligence into products that solve real operational problems for organisations.

My work covers product strategy, enterprise integration, deployment leadership and commercial growth, with a focus on high-volume, customer-facing organisations. I work primarily with banks and other financial institutions, telecommunications providers, technology and infrastructure companies, and public-sector organisations exploring AI for customer service, collections, onboarding, customer reactivation and other operational workflows.

A major part of that responsibility is ensuring that what we build can move beyond a controlled demonstration. Enterprise AI must integrate with existing software, operate reliably under real-world conditions, respect access restrictions and provide a clear record of its actions.

I have also been involved in shaping our approach to voice infrastructure and locally relevant deployment. The objective is not simply to add a conversational interface to an organisation. It is to build agents that can securely connect with business systems and complete authorised workflows.

This intersection of voice, enterprise integration and data control is where I believe BimpeAI can make a meaningful contribution.

What distinguishes BimpeAI from a conventional chatbot provider?

Many chatbots are designed mainly to answer questions from a fixed body of information. Enterprise agents need to do more.

BimpeAI is developing agents that can communicate across voice and messaging channels, connect with operational systems and execute clearly authorised workflows. Depending on the deployment, this could include assisting with bookings, onboarding, collections, customer support, payments or order management.

The distinction is between an AI system that only generates a response and one that can safely participate in a business process.

That requires integration, permission management, observability, escalation procedures and rules governing what the system may do without human intervention. These infrastructure and governance requirements are central to our product approach.

What role should the Nigerian government play in supporting sovereign AI?

Government has two complementary responsibilities: regulation and enablement.

Regulators, including the Central Bank of Nigeria and the Nigeria Data Protection Commission, should provide clear and workable rules concerning data residency, cross-border transfers, AI accountability and customer protection.

At the same time, government should create an environment in which Nigerian technology companies can build and scale. This includes access to computing infrastructure, local cloud capacity, research funding, responsibly developed language datasets, public-private pilot programmes and transparent procurement pathways.

Universities, research institutions, regulators and private companies should also be encouraged to collaborate on applied AI research and shared technical standards.

Regulation without enablement may slow responsible innovation, while enablement without adequate regulation may expose citizens and institutions to avoidable risk. Nigeria needs both.

Where does BimpeAI fit into Nigeria’s broader AI ecosystem?

BimpeAI is building enterprise AI agents that operate across voice and messaging channels, integrate with business systems and support real organisational workflows.

We began by addressing practical business problems, but our work increasingly involves the infrastructure and governance considerations that arise when larger institutions deploy AI.

We are developing voice and workflow agents for Nigerian and African enterprise environments. We are also exploring voice models and deployment architectures that can operate closer to enterprise data, including on-premise and locally controlled environments.

Our objective is to help organisations adopt AI in a way that is useful, reliable, locally relevant and compatible with their governance responsibilities.

For us, sovereign AI is not simply a policy expression. It is becoming a product requirement. If banks, public institutions and large enterprises are to deploy AI at scale, the underlying infrastructure must be trusted, accountable and adapted to the environment in which it operates.

What is the biggest mistake Nigeria could make during the AI era?

The biggest mistake would be to treat artificial intelligence solely as something Nigeria consumes.

Nigeria has a large population, considerable linguistic diversity, a sophisticated financial-services industry, a young technology workforce and major institutional challenges that AI could help address.

If we depend exclusively on imported models and external infrastructure, Nigerian organisations may continually have to adapt their operations to systems that were not designed around their languages, regulations or social context.

Nigeria should participate in the global AI ecosystem while also developing local capabilities in language technology, voice infrastructure, data governance and enterprise deployment.

The opportunity is not merely to use AI tools. It is to build some of the infrastructure, products and standards that will determine how AI works for Nigeria’s economy.

That is how the country can move from being primarily a consumer of artificial intelligence to becoming an important builder and contributor to its future.

About Samuel Adekunle

Samuel Adekunle is the Co-founder of BimpeAI, an enterprise AI infrastructure company developing voice and multichannel agents for high-volume organisations. His work spans enterprise AI deployment, voice infrastructure, systems integration, workflow orchestration, African-language AI and data-conscious deployment models.

Through BimpeAI, he is working to help organisations deploy AI agents that connect with existing business systems, support customer engagement and execute authorised workflows across voice, messaging and digital channels.