Advancing Regtech and Suptech

Case Studies

Case studies from our Public-Private Secondments for Suptech Innovation (PPSSI) workstream.

Join our Private-Public Secondments for Suptech Innovation Workstream

We are now accepting expressions of interest for our 2026 Cohort.

Project BackgroundAcross many emerging economies, financial supervisors face a common dilemma: how to modernize supervisory infrastructure, including the collecting, processing, and acting on regulatory data, without disrupting day-to-day operations or overextending institutional capacity. Mongolia's Financial Regulatory Commission (FRC) is embarking on tackling this challenge. The FRC is the primary regulator for Mongolia's non-banking financial sector, overseeing insurance companies, securities firms, investment funds, and credit institutions. As Mongolia's financial markets grow and digitalize, the FRC faces mounting pressure to supervise an increasingly complex and data-intensive landscape with existing tools and processes that were not designed for this purpose.Recognizing the need for a strategic approach to technology adoption in supervision, the FRC engaged with Regnology through the Public–Private Secondments for SupTech Innovation (PPSSI)programme. The project aimed to draw on Regnology's deep experience in regulatory technology to help the FRC develop a clear, actionable suptech roadmap, grounded in the realities of the Mongolian regulatory context and aligned with international good practice.Project ObjectivesThe collaboration between the FRC and Regnology focused on assessing the current state of supervisory technology at the FRC and identifying a sequenced path toward a more modern, data-driven supervisory function. Specifically, the project sought to:Understand the FRC's existing data infrastructure, reporting workflows, and technology systems;Identify key gaps and pain points in the current supervisory data lifecycle — from collection and validation through to analysis and reporting;Benchmark the FRC's suptech maturity against international peers and good practices;Develop practical, prioritized recommendations across short, medium, and long-term time horizons; andProvide visual, accessible guidance that FRC leadership and technical teams could use to drive forward digital transformation.Solution OverviewMarek Włodarczak, SupTech Subject Matter Expert at Regnology, conducted a structured discovery process with the FRC, including interviewing key staff, reviewing existing systems and data flows, and mapping the end-to-end supervisory data journey. This diagnostic phase revealed important challenges that included a high degree of manual steps in supervisory processes, limited data quality checks, and the use of Excel-based workflows for core tasks. Processes such as risk assessment and complaint handling were also not yet digital, leading to inefficiencies and increased potential for errors.Building on this diagnostic, Regnology developed a SupTech Roadmap for the FRC, providing a structured approach for integrating new technologies into supervisory workflows and prioritizing automated data reporting. This kind of strategic roadmap is particularly important, as financial authorities without a formal suptech strategy tend to report greater implementation hurdles when deploying new technology. The roadmap is structured around a modular suptech architecture, and outlines how components can be introduced incrementally rather than through a single, high-risk transformation. Key elements include:Automated Data Reporting (ADR): A foundational pipeline enabling structured, machine-readable data submissions from regulated entities, reducing manual processing and improving data quality.Supervisory Data Warehouse: A centralized repository integrating data from multiple sources, enabling cross-sector analysis and longitudinal trend monitoring.Risk-Based Supervision Tooling: Analytical modules supporting early warning indicators, peer benchmarking, and sector-level risk dashboards.Integration with National Infrastructure: Guidance on aligning the FRC's data architecture with FINA and the National Data Center, supporting broader government data initiatives.Batchimeg Batbold, Director, Sandbox Unit, Financial Regulatory Commission of Mongolia highlighted the practical value of the project, noting that "as a developing country, we continually face resource constraints – whether human, financial or technological. Having a clear roadmap allows us to prioritize effectively, identify where we can create the greatest supervisory impact and modernize our processes based on the realities of our capacity and needs." Building on this, project secondee Marek Wlodarczak emphasized that addressing these constraints requires a fundamental shift in supervisory approach: “To meet the demands of a complex and rapidly evolving financial landscape, supervisory authorities must transition from traditional, manual oversight to a more agile, data-centric paradigm. We believe this transformation hinges not just on adopting new technology, but on strategically building a strong foundation of data governance and pursuing a phased, value-driven implementation. The principles we have outlined provide a pathway for the FRC to unlock the power of their data, enabling a move from reactive processes to proactive, risk-based supervision and fostering a more resilient financial ecosystem.” Impact and Lessons LearnedThe project generated tangible value by equipping the FRC with a structured, evidence-based framework for suptech investment. FRC Mongolia is now preparing to use these insights to deploy a priority suptech use case before launching a broader suptech solution. Key insights from the engagement include:Context matters: The most effective suptech advice is grounded in the specific regulatory and institutional context of the authority. Generic frameworks need significant adaptation to be actionable for regulators like the FRC.Sequencing is strategic: A phased roadmap that distinguishes short-, medium-, and long-term actions helps regulators make the case for technology investment incrementally, rather than requiring large upfront commitments.Feedback loops improve outcomes: The FRC's detailed mid-project feedback produced a significantly stronger final deliverable. The willingness of both parties to iterate is a hallmark of effective secondments.Data foundations first: The roadmap's emphasis on foundational infrastructure (ADR, data standardization, documentation) before advanced analytics reflects a lesson shared by many supervisors globally: analytics capability is only as strong as the underlying data quality. Next Steps: From Roadmap to ImplementationThe SupTech Roadmap positions the FRC to make informed, sequenced investments in supervisory technology over the coming years. Immediate next steps may include:Sharing the roadmap internally across FRC leadership and technical teams to build alignment on priorities.Engaging with national counterparts to explore data-sharing and integration opportunities.Scoping a pilot for automated data collection in one regulated sector, as a proof-of-concept for the broader ADR architecture.Connecting with development finance institutions and international technical assistance providers that may support implementation. About the Project PartnersFinancial Regulatory Commission of Mongolia (FRC) The FRC is the primary non-banking financial sector regulator in Mongolia, overseeing insurance, securities, investment funds, and credit institutions. The FRC led the project, providing regulatory context, institutional knowledge, and strategic direction to ensure the roadmap reflects Mongolia's supervisory priorities and capacity.FRC Project Contacts: Batchimeg Batbold – Director, Market Research and Development DepartmentJargaltsetseg Batsaikhan - Director, Sandbox Unit Regnology Regnology is a leading international provider of regulatory reporting and SupTech solutions, working with central banks, financial supervisors, and regulated entities across more than 60 countries. Regnology contributed deep technical expertise in supervisory data architecture, regulatory reporting automation, and SupTech implementation to the engagement.Seconded Expert from Regnology:Marek Włodarczak – SupTech Subject Matter Expert Public-Private Secondments for SupTech InnovationCo-led by the World Economic Forum and the Cambridge SupTech Lab, this initiative connects technical experts from the private sector with financial authorities through short-term secondments. The goal is to accelerate innovation in supervision and regulation by solving practical challenges with cutting-edge technology.
Mongolia | Financial services
Building Mongolia's SupTech Future: How Regnology and the Financial Regulatory Commission are Designing a Roadmap for Data-Driven Supervision
Regnology, Mongolia Government
Project BackgroundFor many financial authorities, regulatory guidelines are issued in static PDF documents. As financial systems grow more complex and digital, this approach has become increasingly difficult for supervisors to maintain, and update, and for regulated financial firms to navigate and operationalize. The Bank of Mauritius (BoM) regulates a wide variety of entities, including banks, payment service providers, and FX dealers. BoM’s implementation of a risk-based supervisory framework emphasized significant challenges with this manual approach, particularly as both structured and unstructured information requirements increased. As supervisory expectations become more granular and data-intensive, the reliance on static PDFs strains supervisory capacity, slows updates, and makes it harder to ensure consistency and accuracy across regulatory guidance and supervisory assessments.Currently BoM’s guidelines are in a PDF format, resulting in a bulky manual management process. To update guidelines and rules, BoM accesses a word document copy of the existing guideline, applies the changes and then publishes it on the website as a pdf. This manual process is time-consuming and impacts accuracy, creating a "compliance cost" for both the regulator and the firms it supervises.To address these hurdles, BoM’s Innov8 team launched a Proof of Concept (PoC) to explore the feasibility of "Rules as Code" or converting regulatory requirements into machine-readable formats to streamline supervision and compliance.Urvashi Soobarah, Director of Supervision at the Bank of Mauritius, explains “A key element of our suptech strategy involves migrating from static PDF‑based regulations to structured, machine‑readable rules. This transformation will facilitate the integration of AI/ML‑tools into our supervisory toolkit. Our regulatees may in turn leverage on regtech solutions for compliance management.”Project ObjectivesThrough the Public-Private Secondments for SupTech Innovation initiative, a seconded expert from TSO, a UK-based provider of regulatory technology, collaborated with BoM. The project objective was to deliver a Proof of Concept that demonstrates feasibility and outlines the vision and implementation steps for a working prototype capable of:Automatically converting existing BoM PDF content into machine-readable JSON format.Managing regulatory content within a dedicated content management system (CMS) to allow for easier editing and version control.Demonstrating interoperability by showing how third-party regtech providers could ingest regulatory content directly from the Bank via API.Enhancing transparency by providing the same content in multiple formats (HTML, JSON, and PDF) from a single source.A helpful analogy to solidify understanding would be to think of this project as moving from a printed encyclopedia to Wikipedia with an API. Instead of supervisors and banks manually flipping through static pages (PDFs) and re-typing information, they now have a dynamic, living database. Regulatory changes are made once and instantly reflected across all official publication formats, ensuring consistent interpretation by supervisors, regulated entities, and automated compliance tools. Solution OverviewThe solution outlined in TSO’s PoC was a SaaS-style, web-based platform designed to ingest, transform, and publish regulatory content.1. Content Ingestion & TransformationThe proposed system would allow users to upload PDF guidelines and specify where the main content starts, ensuring the pipeline ignores cover pages or preambles.A custom API would transform the PDF into a tree-hierarchy JSON schema that maps the document down to the paragraph level, complete with unique IDs for tracking and citation.2. Editorial Management via DrupalThe solution would utilize a Drupal CMS hosted on AWS, providing an intuitive "What You See Is What You Get" (WYSIWYG) interface for supervisors to make editorial amendments.When a section is edited, the system would track revisions and allow supervisors to preview changes immediately across all output formats.3. Multimodal OutputFrom a single editorial source, the system would be able to generate a user-focused website (HTML), a machine-readable file (JSON), and a formatted publication (PDF).This ensures that both human readers and automated compliance systems are working from the exact same authoritative data.  Technology StackScalability Potential The diagram below indicates the future development pathways. The PoC illustrates the type of capability required to progress from basic e-governance, through advanced e-governance and into digital first governance (column 1 to column 3). Impact and Lessons LearnedThe PoC successfully demonstrates that digitizing regulations can significantly reshape the supervisory landscape. Key impacts identified include:Operational Efficiency: Automated updates allow for faster revision cycles compared to manual Word-to-PDF workflows.Reduced Compliance Costs: By sharing guidelines via API, firms can ingest rules directly into their internal systems, reducing the "compliance burden."Capacity Building: The digitized format paves the way for internal tools like AI chatbots, which supervisors can use to query regulations or compare them against bank-submitted policy documents.Leapfrogging Development: Urvashi Soobarah emphasized the value of the collaboration, stating, "We really believe that with the assistance through this secondment, we are actually able to leapfrog in this development.”Alan Blanchard, TSO Business Development Director shared “We were delighted to support the Bank of Mauritius Innov8 team in exploring how machine-readable regulation can improve interoperability, accuracy and usability for both regulators and firms. This proof of concept shows how converting unstructured regulatory content into structured, machine-readable formats can remove friction from regulatory change, improve access to authoritative information and ultimately support a more future-ready financial ecosystem.” Next Steps: Toward "Digital-First" GovernanceFollowing the successful PoC, the Bank of Mauritius is pursuing a phased roadmap to move from manual, document-based processes toward digital-first regulatory governance. The next phase focuses on implementing a Minimum Viable Product (MVP) that builds directly on the PoC by operationalising the platform for use by BoM policy and supervisory teams.The MVP will involve setting up a secure, cloud-hosted (AWS) environment that BoM staff can access directly, alongside expanded editorial tools, version control and limited workflow functionality. This will allow supervisors to actively create, amend and publish machine-readable regulatory content, while testing interoperability through APIs and assessing how the system supports both suptech and regtech use cases in day-to-day supervision.In parallel, the MVP will serve as a live testing environment for governance, operational processes and internal capability building, enabling the Bank to evaluate performance, usability and organisational readiness. Insights from this phase will inform decisions on scaling to a full production system, including more advanced workflow automation, broader content coverage and the future integration of AI-enabled supervisory tools.About the Project PartnersTSOTSO (The Stationery Office) is a leading RegTech provider specialized in turning authoritative information into useful, structured data formats.Alan Blanchard – Business Development Director Paul Mansfield – Solution ArchitectBank of MauritiusThe BoM’s Innov8 team and Supervision Department led the coordination, ensuring the envisioned tool aligns with the bank’s digital transformation goals and its commitment to fostering financial innovation. Urvashi Soobarah – Director of SupervisionRanjeet Kallychurn – Senior Analyst Vishnu Nursoo – Bank Officer, Supervision - Licensing & Regulatory Policy UnitPrista Aumeer – Business Intelligence DeveloperPublic-Private Secondments for SupTech InnovationCo-led by the World Economic Forum and the Cambridge SupTech Lab, this initiative connects technical experts from the private sector with financial authorities through short-term secondments. The goal is to accelerate innovation in supervision and regulation by solving practical challenges with cutting-edge technology.  
Mauritius | Financial services, Technology
Digitizing Financial Regulations: How the Bank of Mauritius in Collaboration with TSO is Turning Rules into Code
Bank of Mauritius
OverviewFor central banks and other financial authorities, timely access to consistent, high-quality and well-governed data is critical to effective supervision, risk monitoring, and policy decision-making. As supervisory mandates expand in scope and complexity, authorities increasingly rely on data architectures that support end‑to‑end data ingestion, validation, storage, analytics, and visualisation. A centralized data warehouse (CDW) is a core component of such architectures, enabling the consolidation of heterogeneous data sources into a single environment and supporting advanced, risk‑based supervision. Beyond technical integration, a CDW plays a critical role in strengthening data governance; by centralizing supervisory and operational data, it provides a foundation for consistent data definitions, clear data ownership, standardized quality controls, and auditable data lineages. These elements are essential to ensure data integrity and accountability across the supervisory data lifecycle, and to enable scalable analytics, and the future adoption of advanced suptech tools.The National Bank of Moldova’s (NBM’s) existing IT infrastructure consists of multiple legacy and sector-specific platforms that have evolved over time to support different supervisory and operational functions. While these systems are effective for individual use cases, their limited interoperability makes it challenging to obtain a holistic, cross‑sector view of supervisory data and constrains the ability to perform integrated analysis, cross‑report consistency checks, and forward‑looking risk assessments.In response, NBM has begun exploring the potential development of a centralized data warehouse as part of a broader supervisory data modernization agenda. The objective is to consolidate data from across supervisory, regulatory, and operational systems into a unified data layer that supports consistent access, improved data quality, and enhanced analytical capabilities across the institution.As part of the Public-Private Secondments for SupTech Innovation (PPSSI) initiative, the National Bank of Moldova partnered with a team from DANA to explore how a centralized data warehouse could be designed and developed to support NBM’s supervisory and operational data requirements. The secondment project focused on foundational research, peer learning, and hands-on technical experimentation. DANA’s work was designed as an exploratory effort to inform NBM’s internal decision-making on potential pathways toward a future centralized data architecture. This collaboration reflects DANA’s ongoing commitment to advancing safe tech, and ensuring technology is developed and deployed. Drawing on experience in operating secure, scalable financial infrastructure, DANA contributed technical perspectives on building resilient data architectures that prioritize data integrity, confidentiality, and responsible use.Research and Initial AssessmentAt the outset of the secondment, the DANA team conducted a research phase focused on assessing the current state of NBM’s data environment and reviewing possible paths toward the development of a CDW. The secondment team reviewed and synthesized relevant internal documentation, conducted research on global supervisory data architectures, and participated in structured peer learning with other financial authorities that have implemented similar solutions.NBM’s supervisory and operational data environment is characterised by fragmented, siloed systems with limited integration across platforms. Based on available documentation and discussions with NBM counterparts, the DANA team gained an initial understanding of NBM’s existing data processes. Through this work, the team identified several key challenges, including a heavy reliance on manual data extraction, calculation, and reporting; limited interoperability between core supervisory systems; and the absence of integrated monitoring capabilities. These factors highlighted opportunities to strengthen data consistency and enhance analytical capabilities through a more integrated data architecture.In parallel, the DANA team reviewed centralized data warehouse approaches adopted by other central banks and supervisory authorities and engaged in structured learning sessions with peers that have implemented centralized data warehouse initiatives. These activities helped inform early thinking on architectural options that could be adapted to NBM’s institutional, regulatory, and technical context.Solution ApproachFollowing an assessment of NBM’s existing data environment and operational needs, the DANA team recommended a modular technology stack to support a future data warehouse and analytics capability:ETL and data integration: TalendData warehouse platform: ClouderaData analytics and visualisation: TableauThe DANA team provided initial licensing/hardware cost estimates for the proposed stack and discussed feasibility with the NBM team. From there, NBM obtained trial licences for the recommended software, and the teams jointly set up an on‑premise sandbox environment using an NBM data sample.Within this controlled sandbox environment, the DANA team:Configured the Talend–Cloudera–Tableau stack on NBM’s infrastructure;Conducted an end‑to‑end supervisory data use case, from ingestion and transformation to analytics and visualisation;Demonstrated how automated workflows could replace manual calculations and reporting steps; andDeveloped step‑by‑step technical documentation to support internal knowledge transfer and future experimentation.The project concluded with a live demonstration using sample data, allowing the NBM team to ask questions and discuss practical considerations for future scaling.In reflecting on the project, the DANA technology team remarked “Rather than delivering a fixed solution, we created a safe and trusted space for experimentation, enabling supervisory teams to explore modern data workflows while prioritizing safety, data protection, and governance. Through hands-on collaboration in a secure sandbox environment, we translated architectural concepts into practical processes and strengthened shared understanding of responsible data management, including ownership, quality, and operational resilience. This learning-driven approach reflects DANA’s commitment to building trusted, secure, and future-ready supervisory systems.” Outcomes and ValueThe secondment delivered several tangible outcomes for NBM:Greater insight into the technical and organisational requirements for a centralized data warehouse;Practical insights into how modern ETL, data management, and analytics tools could reduce manual effort and improve data consistency;A clearer understanding of data integration challenges and dependencies across existing systems; andA sandbox environment and technical documentation that NBM can reuse for continued internal capacity building.Reflecting on the collaboration, Mihai Ursu, Chief Digital Officer, Department of Information Technology, at NBM noted “For an institution at an early stage of its data architecture transformation, having access to hands-on technical expertise and the experience of peer central banks through this initiative proved very valuable. The sandbox environment we built together now serves as both a proof of concept and a practical reference point for our upcoming pilot project.”The project equipped NBM with the knowledge and practical experience needed to make informed choices about the next phase of its data transformation journey.Next StepsBuilding on the outcomes of the secondment, NBM plans to conduct a pilot in 2026 to further experiment with the DANA team’s recommended software stack and architectural approaches. Lessons learned from the pilot will inform a broader supervisory data transformation and the development of a formal data governance framework to be implemented alongside any future centralized data warehouse.Project PartnersDANADANA is a digital wallet-based technology financial company that provides payment platforms and financial services in Indonesia. Since its launch in December 2018, DANA has now reached more than 200 million users in Indonesia. DANA provides complete digital transaction solutions that are safe, easy and convenient for users, merchants and financial institutions. The DANA solution, developed by leading financial technology engineers in Indonesia, aims to advance and accelerate financial literacy and inclusion in Indonesia. Apart from empowering digital finance, DANA also seeks to encourage Indonesian society through impactful social initiatives that focus on empowering women and preserving the environment.Seconded experts from DANA Indonesia:Bimandika Hasanah, IDC System Engineer LeadWinadi Wiratama, Senior Software Development EngineerAlfredo Juan Pratama, Senior Software Development Engineer in TestAhmad Irfaan Hibatullah, Software Development EngineerNational Bank of Moldova (NBM)The National Bank of Moldova is the central bank of the Republic of Moldova. NBM is responsible for conducting monetary policy, safeguarding price and financial stability, and regulating and supervising the financial sector. Its supervisory mandate covers banks, non‑bank credit organisations, savings and loan associations, payment service providers, and e‑money issuers, as well as oversight responsibilities related to financial market infrastructure and, more recently, insurance and non‑bank financial sectors.NBM plays a central role in prudential supervision, macroprudential oversight, and the implementation of international regulatory standards. In recent years, the institution has prioritised the modernisation of its supervisory processes, including the adoption of suptech and data‑driven approaches, to support risk‑based supervision across an increasingly diverse financial system.NBM team included:Efimia Martinescu - Head of Strategic Planning & Management, Strategy & HR DepartmentEugeniu Aftene - Director, Statistics & Data DepartmentIon Pascaru - Head of ICT Organisation & Management, IT DepartmentMihai Ursu - Chief Digital Officer, IT DepartmentValeriu Iordachi - Head of ICT Infrastructure, IT DepartmentVictor Ababii - Head of Data Collection & Processing, Statistics & Data DepartmentPublic-Private Secondments for SupTech InnovationCo-led by the World Economic Forum and the Cambridge SupTech Lab, this initiative connects technical experts from the private sector with financial authorities through short-term secondments. The goal is to accelerate innovation in supervision and regulation by solving practical challenges with cutting-edge technology.
Moldova | Financial services, Government, Technology
DANA Contributes to Foundational Exploration of a Centralized Data Warehouse at the National Bank of Moldova
DANA, National Bank of Moldova
Project BackgroundConsumer protection and financial education are core mandates for central banks worldwide. In many markets, consumers struggle to make informed choices when selecting financial products. Fee structures are complex, terminology varies across institutions, and making comparisons between products and providers can be difficult. As a result, consumers may select suboptimal products, incur avoidable costs or spend an inordinate amount of time trying to compare options, breeding mistrust and frustration within the financial system.Recognizing these challenges, a central bank in the Middle East is in the process of building a real-time tool that captures fee and product information from all domestic financial institutions and displays it in an easy to digest format for consumers. This tool will allow consumers to easily compare different financial products across institutions and empower consumers to make an informed decision.To support this effort, a team from DANA, a digital wallet fintech company, conducted targeted research through a secondment project under the Private-Public Secondments for SupTech Innovation (PPSSI) program. The objective of the secondment was to assess existing global fee comparison tools to help inform the design of the central bank’s future solution.Research and OutcomesThe team from DANA reviewed nine different fee comparison tools, five from financial authorities, and four from the private sector. The team assessed the user experience, governance mechanisms, and data update processes of each tool to understand how accuracy, transparency, and timeliness were maintained over time.The secondment team consolidated the findings into a final document, including an analysis of each tool’s strengths and weaknesses, as well as a summary of key takeaways the central bank team should consider as they develop the fee comparability tool.Based on the research and survey conducted, the team came up with the following key aspects of a user-friendly fee comparison tool:Simplicity: Intuitive navigation, clear language, and avoidance of information overload.Transparency: Clear disclosure of data sources, update frequency, and any limitations of the tool.Guidance: Contextual tips, explanations, and live support integrated throughout the user journey.Personalization: Default scenarios for common user needs, with options for advanced customization.Actionability: Downloadable and printable results, along with direct calls-to-action such as “Contact this bank.”Inclusive Design: Multiple language options, accessible reading levels, and features supporting users with visual impairments.One of the strongest insights from the research was that consumer-facing transparency tools are only as effective as the clarity and consistency of the underlying data. Beyond interface design, sustainable impact depends on well-defined update processes, clear data ownership, and governance structures that ensure information remains accurate and trustworthy over time. Effective fee transparency is not only about presenting information, but about enabling consumers to make confident, informed decisions. Jason Widjaja, Lead UX Researcher at DANA and secondee on this project noted “By studying a range of global tools, we saw how thoughtful design choices, from clear language to inclusive access, can significantly improve financial understanding and reduce information asymmetry in the market.”Next StepsBuilding on these insights, the central bank is well-positioned to move forward in the fee-comparability tool development process, and plans to do so. By leveraging global best practices and lessons learned from the DANA team’s research, the central bank can avoid common pitfalls and ensure that consumers drive the most benefit from the final product. About the Project PartnersDANA DANA is a digital wallet-based technology financial company that provides payment platforms and financial services in Indonesia. Since its launch in December 2018, DANA has now reached more than 200 million users in Indonesia. DANA provides complete digital transaction solutions that are safe, easy and convenient for users, merchants and financial institutions. The DANA solution, developed by leading financial technology engineers in Indonesia, aims to advance and accelerate financial literacy and inclusion in Indonesia. Apart from empowering digital finance, DANA also seeks to encourage Indonesian society through impactful social initiatives that focus on empowering women and preserving the environment.  Seconded expert from DANA:Jason Widjaja, Lead UX Researcher Public-Private Secondments for SupTech InnovationCo-led by the World Economic Forum and the Cambridge SupTech Lab, this initiative connects technical experts from the private sector with financial authorities through short-term secondments. The goal is to accelerate innovation in supervision and regulation by solving practical challenges with cutting-edge technology. 
Middle East & North Africa | Financial services
DANA Conducts Comparative Research on Consumer Fee Transparency Tools for a Central Bank in the Middle East
DANA
Project BackgroundAcross many financial systems, consumer complaints represent one of the most critical but under-leveraged sources of supervisory intelligence. Central banks receive thousands of complaints from the public each year. However, these are often submitted in unstructured formats: handwritten letters, scanned images, PDFs, and forms in multiple languages. Manually reviewing, classifying, and acting upon such data is time-consuming and prone to inconsistencies.To overcome this challenge, a central bank in the Middle East decided to explore the use of AI technologies to automate the extraction and structuring of complaint data. Through the Public-Private Secondments for SupTech Innovation initiative, the World Economic Forum and the University of Cambridge SupTech Lab brought in a dedicated seconded team from Abhi, a fintech infrastructure company based in Pakistan. The team included experts in OCR, natural language processing, and large language models, who collaborated directly with the central bank's supervision and IT departments.Aligned with the Bank's broader digital transformation strategy, the project’s objective was to co-design a proof of concept capable of:Extracting text from a variety of unstructured document formats.Analysing the extracted text to identify key complaint themes and metadata.Converting free-text complaints into structured data that could be analyzed or integrated into existing systems.Achieving high accuracy across multilingual inputs (Arabic and English).Ensuring that the solution could be deployed on-premise, with no reliance on cloud infrastructure.Solution OverviewThe proof of concept developed by Abhi was a modular, AI-driven pipeline capable of ingesting and analyzing consumer complaints in multiple formats and languages. The solution design combined several technologies into a single architecture:Optical Character Recognition (OCR)The solution design leveraged a custom-trained OCR engine (Surya OCR) that performed well on scanned Arabic text, including noisy or handwritten documents, achieving 91.6% character-level accuracy for low-resolution and poorly scanned documents.Pre-processing steps such as skew correction, noise filtering, and contrast and resolution enhancement were applied to optimize OCR outputs and improve reliability, resulting in a 13% improvement in OCR accuracy.The system could handle various layouts, such as bank forms, handwritten notes, and multi-column PDFs.Multilingual Text Extraction and PreprocessingOnce text was extracted, it was normalized, language-tagged (Arabic or English), and segmented into tokens for further processing.Auto-orientation and layout parsing steps ensured clean, language-aware inputs to downstream NLP modules.The architecture supported a dual-language NLP pipeline to ensure semantic consistency across Arabic and English complaints.Keyword and Entity RecognitionA combination of rule-based matching and trained Named Entity Recognition (NER) models, such as Qwen-72B, was used to identify entities such as:Financial product types (loans, cards, savings)Complaint themes (fraud, delay, miscommunication)Institution names or locationsA taxonomy of keywords was developed by the central bank and used to standardize output fields.Large Language Model (LLM) IntegrationTo enhance prediction accuracy and interpretability, Abhi integrated a Retrieval-Augmented Generation (RAG) mechanism and Labeling for Retrieval Augmentation (LfRA) into the system.The system retrieved historical resolution examples using ChromaDB and learned from human-corrected outputs to refine future predictions.This LLM module was fine-tuned on complaint-style text and used to:Disambiguate noisy OCR outputIdentify implied intent and complaint toneImprove field tagging through context-aware reasoningOutput StructuringAll extracted and interpreted data was transformed into a structured JSON format that mapped each complaint to fields such as:Date, product type, nature of complaint, involved institution, suggested next step (optional).The system maintained traceability and auditability by linking each output field back to its source text span.Technology StackOCR EnginesSurya OCR: Main engine optimized for handwritten Arabic text.Tesseract: Backup engine for typed Arabic text.            Named Entity Recognition (NER)Qwen 2.5 - 72B: Accessed via OpenRouter for contextual entity recognition and improved NLP accuracy. Extracts names, dates, complaint types, and other key fields.Sentiment AnalysisRule-enhanced model: Flags complaints with strong negative sentiment or urgency indicators.Backend FrameworkDjango: Handles complaint-processing logic and exposes REST APIs for integration.Evaluation and AccuracyThe solution design was tested on a diverse set of real complaint documents provided by the central bank, including samples categorized as A (complex), B (moderate), and C (simple). Abhi conducted side-by-side accuracy benchmarking with and without the LLM enhancement.The team also conducted manual validation and native Arabic experts alongside central bank staff, who verified that extracted data aligned with the actual complaint contents.Open-Source Delivery and On-Premise ReadinessA key achievement of the project was that all outputs were delivered as open-source, removing concerns about vendor lock-in. This included the full OCR and NLP pipeline, keyword taxonomies, model weights (where licensing allowed), and user documentation.Recognizing the requirements for data privacy and infrastructure control, Abhi also provided:A detailed on-premise deployment plan, including system architecture diagrams and VM requirements.Containerized services (via Docker) to ease local deployment.Instructions for integrating with the Bank's existing IT stack and secure environments.Recommendations for future updates and retraining based on evolving complaint patterns.This ensured that the solution could be managed entirely within the Bank's infrastructure — aligning with their compliance and data protection policies.Impact and Lessons LearnedThis project represents a strong example of how targeted technical secondments can generate solution designs that address concrete supervisory challenges. Key impacts include:Operational Efficiency: Complaints that once required manual reading and tagging can now be structured with high accuracy in seconds.Capacity Building: The internal supervision and IT teams were directly involved throughout, building familiarity with modern AI tooling.Technology Transfer: All code, models, and documentation were transferred in open format, enabling future upgrades.Replicability: The modular design means that similar authorities can adopt the approach in other jurisdictions.Next Steps: Toward Resolution IntelligenceFollowing the proof of concept’s insight, the central bank is now exploring operational deployment and extending the system to link structured complaints to predefined resolution actions, taking into account urgency, content, and regulatory significance. This could enable faster triage, consistent response strategies, and improved consumer outcomes.Abhi proposed integrating lightweight AI agents trained on tagged resolution pathways, the system could dynamically recommend actions based on complaint attributes. These agents could operate on top of the structured JSON output already produced, making the extension both cost-effective and immediately impactful.This roadmap would allow the Bank to move from reactive complaint handling to intelligent supervisory resolution, aligning with its broader digital transformation goals.About the Project PartnersAbhi Private LimitedAbhi is a leading fintech infrastructure company, headquartered in Pakistan, specializing in real-time payments and embedded finance. The seconded team included:Ejaz Anwer – Technical Lead & Deployment ArchitectAhmad Khalid Butt – Solution Designer, NLP & Machine Learning EngineerAbdul Bari – Arabic Language ExpertPublic-Private Secondments for SupTech InnovationCo-led by the Cambridge SupTech Lab and the World Economic Forum, this initiative connects technical experts from the private sector with financial authorities through short-term secondments. The goal is to accelerate innovation in supervision and regulation by solving practical challenges with cutting-edge technology.
Middle East & North Africa | Financial services
Decoding Complaints at Scale: How Abhi Unlocked Value from Unstructured Consumer Data for a central bank in the Middle East
Abhi

Advancing Regtech and Suptech

A World Economic Forum Initiative
© 2026 All rights reserved
Privacy Policy & Terms of Service