Regtech and suptech in central banks 2026 case studies: Part one – research and investment priorities

Central Banking’s regtech and suptech interviews are in‑depth exploration of the pioneering work at central banks and supervisory authorities.

We spoke to teams at the Central Bank of Brazil (BCB), Qatar Financial Centre Regulatory Authority (QFCRA), National Bank of Georgia and Peru’s Superintendency of Banking, Insurance and Private Pension Fund Administrators (SBS).

Officials tell Central Banking how they are using cutting-edge technology to monitor a fast-paced fintech landscape. As well as insights into data and infrastructure initiatives that allow predictive modelling, in this first part of a three‑part series, supervisory team leaders talk about their investment and research priorities and how these are evolving.

For example, the BCB has created a corporate data warehouse in order to replace several departmental calculators that were at capacity, and has plans for an institutional cloud platform for generative artificial intelligence (GenAI). An innovation and technology committee was created to manage resources more efficiently.

At the SBS in Peru, The data and analytics department was formally established in 2025 and a re-evaluation of on-premise information collection systems is under way, aimed at enhancing its technological architecture. A competency-based learning approach has been adopted, identifying the skills required at each stage of the data lifecycle, defining the corresponding data roles and designing training pathways for each one.

At the QFCRA, the financial analysis and innovation team sits within the supervisory division and develops suptech solutions in close collaboration with the IT department. Under the supervision and authorisation division’s 2025–30 strategy, a digitalised financial ecosystem is a key aim, encompassing support for regtech in reporting and “deeper synergies” with the team’s suptech capabilities.

Additionally, suptech has the potential to strengthen the work of other departments. In Peru, anti-money laundering (AML) and combatting the financing of terrorism (CFT) work of the Financial Intelligence Unit has been supported by link analysis tools that integrate large volumes of data, identify hidden patterns and visualise relationships between entities and transactions.

Meanwhile, the National Bank of Georgia has established a Report Lab platform and, as well as more advanced machine learning solutions for systemic risk detection, is exploring blockchain-based solutions for reporting to further enhance regulatory data transparency.

All the teams we spoke to face the challenge of handling multi-source confidential data. Investment prioritises integrated systems that can meet multiple supervisory needs in a unified way. Staff training in data science and AI is also a priority. Meanwhile, AI governance issues, such as ensuring algorithm explainability, are also weighted, as they may imply additional investment in AI audit and compliance tools.

The regtech and suptech evolution reflects a transition from periodic, manual processes toward continuous, automated oversight capabilities. Alongside this digital transformation, Central Banking hopes this interview series contributes to essential knowledge sharing among authorities, monitoring rapidly changing and complex financial relationships across the world.

What are your institution’s key regulatory and supervisory research and investment priorities, particularly in areas such as data analytics, risk modelling and regulatory reporting?

Tatia Tsiklauri, National Bank of Georgia: The National Bank of Georgia has identified several strategic investment priorities to enhance our supervisory capabilities. In data analytics, we are focusing on building an integrated data platform that consolidates information from various sources – regulatory reports, market data, consumer complaints and macroeconomic indicators – to provide a comprehensive view of the financial ecosystem. We are expanding our team’s capabilities with Python and advanced statistical tools while continuing to leverage  Microsoft’s Power BI for accessible visualisation tools for leadership and supervisory teams. We have already implemented XBRL standards for financial institution reporting, which has significantly improved data comparability and analysis capabilities. We’re also now exploring blockchain-based solutions for certain reporting streams to further enhance transparency and immutability of regulatory data.

What is really important is that we are at this point where central banks want to move forward – they want to leverage modernisation and innovation to achieve an environment that is more streamlined, automated and resilient

Tatia Tsiklauri

Aristides Andrade Cavalcante Neto, Central Bank of Brazil (BCB): The BCB’s main priorities in terms of regulatory and supervisory research and investments include intensifying the use of advanced data analytics and AI in supervision, improving risk models by incorporating climate risks and new factors, and modernising regulatory reporting and supervisory technological infrastructure.

There is a strong focus on suptech and data analysis. The BCB recognises that innovative technology solutions applied to supervision allow for increased effectiveness without a proportional increase in resources, which is why it has conducted dedicated programmes, such as S-Lab and suptech projects, to experiment with machine learning, AI, train personnel and develop new tools for supervision processes.

This strategic priority translates into initiatives such as developing algorithms for automated information triage and creating an institutional suptech strategy. Internal studies align with best practices suggested by the Financial Stability Institute of the Bank for International Settlements, seeking to clearly articulate objectives, data needs and an action plan to integrate suptech into supervisory routines. Investing in applied data science and AI is a priority – which involves both training personnel, including data scientists and IT analysts, and acquiring tools and developing pilot use cases in the area of prudential and conduct supervision.

Another important priority is the improvement of risk models and supervisory methodologies. The BCB has historically maintained active research in macroprudential stress test models, analysis of financial interconnectivity networks and crisis prediction. In recent years, new priority topics have emerged, such as climate and environmental, social, and governance (ESG) risks, cyber risks and innovative financial assets. The BCB has joined international networks such as the Network for Greening the Financial System and participates in studies to develop methodologies for quantifying climate risks for banks. This guides investments in climate scenario models and the integration of these factors into traditional risk processes. Similarly, for fintechs and crypto assets, there are analytical efforts to understand their prudential implications and whether existing supervisory tools are adequate or need to evolve. In parallel, the BCB continues to refine classic credit, market and liquidity risk models, incorporating modern techniques.

Andrei Cardoso Vanderlei, BCB: Our key regulatory and supervisory research and investment priorities are centred on deep AI integration in the foundational development of human capital and technological infrastructure. We view the incorporation of AI into our supervisory processes as a fundamental strategic ally for continuously expanding our capacity of supervision.

For instance, in risk modelling, we are prioritising strategic investments in AI applications for specific, evolving risk categories, such as ESG risks and cyber risks, utilising AI and natural language processing (NLP) to monitor and analyse financial institutions’ risk management policies. Reinforcing our existing regulatory framework for model risk management remains a high priority.

We’ve made consistent and significant investments in the technical training of our staff, particularly in supervision and technology. Data science and AI were the most relevant training for supervisors in 2025, indicating a strategic investment in developing an AI-ready workforce. A key infrastructure investment is the launch of an institutional cloud platform designed to enable our departments to develop and test secure GenAI solutions with proper governance.

In risk modelling, we are prioritising strategic investments in AI applications for specific, evolving risk categories, such as ESG risks and cyber risks, utilising AI and NLP to monitor and analyse financial institutions’ risk management policies

Andrei Cardoso Vanderlei

Superintendency of Banking, Insurance and Private Pension Fund Administrators (SBS), Peru: The Institutional Strategic Plan 2026–2030 sets out a strategic objective aimed at guiding the institution through an orderly digital transformation, positioning data as a strategic asset, fostering innovation, strengthening capacities in data analytics, and promoting the use of emerging technologies to optimise regulatory and supervisory processes.

Among the key initiatives is the review and update of the institutional technological framework, with the goal of facilitating scalability and enabling the adoption of cutting-edge technologies. In this regard, the institution is evaluating the contracting of cloud service platforms, and has therefore engaged with various providers to assess innovative solutions available in the market.

Additionally, a re-evaluation of on-premises information collection systems is under way to identify opportunities for improvement and to design a work programme aimed at enhancing their technological architecture.

In the field of data analytics, the data and analytics department was formally established in 2025. This department is responsible for defining guidelines for data management and actively participates in the implementation of initiatives related to structured data and data lifecycle management, in close co-ordination with supervisory departments.

Furthermore, a competency-based learning approach has been adopted, identifying the skills required at each stage of the data lifecycle, defining the corresponding data roles and designing training pathways for each one. This ensures the progressive development of the competencies needed to strengthen institutional analytical capacity.

In line with this approach, budget allocations have been made for external training and for acquiring licences for platforms such as Datacamp and Coursera. This effort is complemented by an internal training programme delivered by institutional experts, which includes practical courses based on SBS databases and real supervisory case studies. To reinforce this strategy, a ‘train the trainers’ programme has also been implemented, aimed at consolidating internal expertise and ensuring the long-term sustainability of the learning model.

What are your institution’s key budgetary considerations when exploring investments in supervisory technology, with a particular focus on the potential for integrating and streamlining workflows and developing a more efficient supervisory infrastructure?

Ewald Müller, Qatar Financial Centre Regulatory Authority (QFCRA): That is probably still one of the biggest challenges. I always laugh at the fact that one of the core principles, both in banking and in insurance, is that the supervisor or regulator should be financially independent. So, in terms of budgetary considerations, the top down meeting the bottom up is a massive challenge. It is alleviated by the fact that some of these initiatives have been undertaken at the state level and then made available to us, so that does cut out costs. So, if you talk about budgetary considerations, I think the major consideration for us is leverage. Qatar is a small country, a very rich country, but it is important for us to pool our resources, to share infrastructure, and that is also where the top down tries to almost enforce that. So that is a good thing, in the sense that you shouldn’t all be going on your own journey. We should be trying to co-ordinate this journey, and we are still not doing it particularly well, but it is better than it was. So technology is our primary focus, but definitely, being able to leverage off a shared platform is critically important for the success of that. Because there might be a lot of money, but we do not have an unlimited budget. That is just the way life works.

Technology is our primary focus, but definitely, being able to leverage off a shared platform is critically important

Ewald Müller

Aristides Andrade Cavalcante Neto, BCB: When evaluating investments in supervisory technology, the BCB carefully considers costs versus benefits, opportunities for process integration and the impact on the operational efficiency of the supervisory infrastructure.

In terms of efficient resource allocation and cost-benefit analysis, like any public institution, the central bank operates with a limited budget and needs to justify expenditures. Thus, suptech projects undergo rigorous feasibility analysis. Priority is given to initiatives with high potential for efficiency gains or risk mitigation, so that the expected benefit – reduction of manual labour hours, prevention of costly crises – clearly outweighs the cost of development or acquisition. Often, this involves opting for mature market solutions instead of developing internally from scratch, if available, to save time and money. This lesson was learned from initial pilots: not all problems require proprietary AI or ground-up construction – if a ready-made tool exists, it may be more economical to license it. At the same time, in critical areas where suitable commercial software is not available, for example to integrate multi-source confidential data, investment in internal development is justified. The BCB has also been leveraging cloud computing and shared multi-cloud infrastructures, which can accelerate deliveries with less need for investment in proprietary data centres.

The budget allocated to suptech prioritises scalable and unifying solutions, avoiding funding duplicated systems or technological islands. Instead, budgetary preference falls on integrated platforms or systems that can meet multiple supervisory needs in a unified way. For example, a single analytical solution that serves different departments – banks, co-operatives, capital markets – may be more costly initially, but it reduces silo maintenance expenses and future integration costs. This integrated vision guided the proposal to create a corporate data warehouse for monitoring, in order to replace several departmental ‘calculators’ that were at their capacity limit.

An aspect often accounted for in investment planning is the need to train teams and ensure the sustenance of new tools. There is no point in acquiring state-of-the‑art software if there is a lack of trained personnel to operate or maintain it. Thus, the BCB includes in suptech projects budget allocations for technical training and hiring specialists such as data scientists and developers, as well as partnerships with universities and other authorities for knowledge exchange. In addition, the longevity of the solution weighs on the decision: choosing open technologies with broad support avoids unforeseen future expenses. For example, if the BCB adopts a proprietary system, the cost of continuous licences and dependence on the vendor is considered against opting for an open-source solution with an active community, which can be maintained internally with less cost. AI governance issues, such as ensuring algorithm explainability, are also weighted, as they may imply additional investment in AI audit and compliance tools.

In more general terms, budgetary constraints and public procurement processes, which are not always agile, demand creativity from the BCB to do more with less in the suptech area. We seek to improve our supervisory processes by balancing innovation and prudence, investing in what brings clear gains – quality, agility, precision – and underpins the future of supervision, ensuring integration and adequate support, while avoiding superfluous expenses or unsustainable solutions. In this regard, the innovation and technology committee was created to establish guidelines, centralise suptech demands, promote innovation and identify synergies to manage resources more efficiently.

Budgetary preference falls on integrated platforms or systems that can meet multiple supervisory needs in a unified way. For example, a single analytical solution that serves different departments – banks, co-operatives, capital markets – may be more costly initially, but it reduces silo maintenance expenses and future integration costs

Aristides Andrade Cavalcante Neto

Andrei Cardoso Vanderlei, BCB: AI’s ability to accelerate and execute repetitive and voluminous tasks, such as document processing, data reconciliation and report generation, with much more precision, directly translates into effort savings. Concrete examples, such as Axis’s capacity to reduce a year of manual tasks to a few days, demonstrate tangible returns on investment by freeing up valuable human resources for higher-value activities.

Budgetary allocations are clearly prioritising foundational infrastructure, exemplified by the planned institutional cloud platform for GenAI. While initially a significant investment, this move towards a scalable, secure, and centralised environment for AI development, and deployment offers long-term efficiencies, reduces redundancy and ensures consistent governance across various AI initiatives. This represents a strategic shift to a more robust, future-proof infrastructure. Our consistent investment in the technical training of our supervision staff highlights that human expertise is a critical budgetary consideration alongside technology. This investment ensures that our workforce possesses the necessary skills to effectively leverage and manage advanced suptech solutions.

What are the pros and cons of off-the-shelf solutions?

SBS, Peru: With respect to market conduct supervision, this type of solution has proven highly useful in optimising supervisory processes. It enables the integration of various data sources – such as complaints, reports and even social media mentions – to identify behavioural patterns and detect early warning signals that require supervisory action.

Through data analytics supported by AI, we can rapidly classify consumer insights into specific supervisory themes, allowing us to anticipate risks, take timely actions and strengthen financial consumer protection.

However, it is important to note that these have not been purely off-the-shelf solutions. Although some tools were procured from third parties or built using publicly available libraries, they required extensive customisation and workflow adaptation beyond basic configuration. This has involved significant investment of time and human resources before such tools could be fully integrated into supervisory processes.

Additionally, these tools have also helped strengthen the work of the Financial Intelligence Unit, particularly in the analysis of AML/CFT cases, through the use of link analysis tools that integrate large volumes of data, identify hidden patterns and visualise relationships between entities and transactions.

Regarding the limitations of such solutions, one of the main challenges is the need for close collaboration between supervisors and developers to design bespoke models that can capture the complexity of supervisory processes. Other challenges include vendor dependency and constraints on system flexibility when new functionalities are required.

One of the main challenges is the need for close collaboration between supervisors and developers to design bespoke models that can capture the complexity of supervisory processes

SBS, Peru

How are your regtech and suptech operations, capacity and costs evolving?

Perttu Korhonen, QFCRA: Our financial analysis and innovation (FAI) team sits within the supervisory division and develops supervisory technology solutions in close collaboration with our IT department. While IT retains platform administration, since 2020 FAI has served as the content administrator for all artefacts developed and shared on our suptech platform, including regulatory best interest. Today, just under 10% of divisional staff are in FAI, and we expect this to rise toward 15% as demand for data-driven supervision grows. We combine supervisory domain expertise with data science, machine learning and analytics skills – crucial for solutions that are technically robust and supervisory relevant. With the growing importance of technology-related risks in our regulated entities, including cyber and AI risks, FAI has become an increasingly important expert partner to the supervisory teams.

Costs and capacity are evolving in tandem. By automating data pipelines and deploying workflows, we are reducing manual effort and redirecting supervisory time to higher-value analysis. Our migration to KNIME Business Hub will improve scalability, collaboration and time-to-value while lowering maintenance overhead versus legacy tooling; these efficiency gains help keep overall costs stable as capabilities expand. We are integrating with government-negotiated Microsoft and Google cloud services to meet data residency requirements and access advanced analytics.

Under the supervision and authorisation division’s 2025–30 strategy, a ‘digitalised financial ecosystem’ is a key outcome, encompassing support for regtech in reporting and compliance and deeper synergies with our suptech capabilities. We are actively engaging with regtech vendors via market scans and proofs of concept. This creates two-way value: products that address our internal use cases move faster toward adoption, and vendors gain structured, non-confidential insight into our supervisory expectations and the state’s regulated sector within clear data protection and confidentiality boundaries.

Costs and capacity are evolving in tandem. By automating data pipelines and deploying workflows, we are reducing manual effort and redirecting supervisory time to higher-value analysis

Perttu Korhonen

SBS, Peru: For market conduct supervision, one of the current solutions follows a hybrid model, developed jointly with an external provider through customised analytics on consumer-friction data. This involved strong internal participation for model training and adaptation to supervisory objectives.

Other ongoing developments are being designed or implemented in-house, using publicly available tools such as web-scraping, web-data analytics, intelligent audio transcription with speaker identification and topic detection for supervisory purposes, and regulatory-document traceability tools related to market-conduct topics.

The development of these solutions has created a growing need to expand AI expertise within the team, leading to the recruitment of specialised profiles and the upskilling of existing staff through both institutional initiatives and individual training. However, a current limitation lies in the available processing capacity, requiring further investment in infrastructure to continue advancing these projects.

In the area of AML/CFT supervision, significant progress has been made in automation and digitisation of continuous supervisory processes – particularly in ongoing transaction monitoring, trend analysis of suspicious-transaction reports and pattern detection. Analytical capacity has increased through the development of models that allow for the analysis of larger and more diverse datasets.

In terms of infrastructure investment, especially for AI, machine learning and automation projects, the AML/CFT area has recently incorporated technologies with enhanced processing capabilities. Nevertheless, the natural obsolescence of such assets and the emergence of new hardware components make regular renewal necessary.

Finally, regarding the SBS solvency stress-testing model, efforts are ongoing to enhance the retail probability of default satellite models by leveraging granular, debtor-level information and applying machine learning techniques to improve the model’s predictive power.

In terms of infrastructure investment, especially for AI, machine learning and automation projects, the AML/CFT area has recently incorporated technologies with enhanced processing capabilities

SBS, Peru

Tatia Tsiklauri, National Bank of Georgia: Our regtech and suptech operations are experiencing significant evolution across multiple dimensions. We have established a foundation with standardised data collection systems using XBRL and our Report Lab platform, which has already demonstrated measurable improvements in data quality and processing efficiency. Our capacity is expanding through strategic investments in our integrated data platform that consolidates regulatory reports, market data, consumer complaints and macroeconomic indicators. We are actively building our team’s capabilities with Python and advanced statistical tools while maintaining our Power BI infrastructure for accessible visualisation.

Current implementations include our regulatory chatbot using NLP, web-scraping tools with text analytics for market monitoring, and developing anomaly detection models. As well as blockchain-based solutions for certain reporting streams, we are exploring more advanced machine learning solutions for systemic risk detection.

The evolution reflects a transition from periodic, manual processes toward continuous, automated oversight capabilities, particularly important for monitoring the rapidly evolving fintech landscape in Georgia.

Sponsored content