BankTech Day 2025 – Case Study: UIB Builds its Data & AI Foundation with Alteryx and Prime Analytics

Presented by Mohamed Bessa (Director of Data and Projects – UIB), William Rank (Director Channel Sales – Alteryx), and Skander Mezghani (CEO – Prime Analytics), the workshop titled « How to Build a Banking Data & AI Foundation and Operationalize It Across Business Use Cases » brought together an audience of Data, IT, and Risk departments from the banking and financial sector.‍The program included: real-world feedback on the design of UIB's Smart Data Hub, implementation of an Alteryx automation factory, first deployed use cases (behavioral segmentation, corporate internal rating, IFRS models), model evaluation and monitoring discipline, and 2026 outlook.

Mohamed Bessa, Director of Data and Projects at UIB, opened the session by presenting the bank and the project context.‍A subsidiary of Société Générale Group, UIB is a Tunisian bank established nearly 60 years ago, which today has 145 branches and approximately 900,000 customers.

‍

"It's a 60-year-old bank. That means a lot of legacy systems, a lot of heterogeneous data, and a huge amount of code — Python, R. We had to manage all of that: volume, heterogeneity, and an absolute need to reduce time to market."

‍

The challenge came down to three words: volume, heterogeneity, acceleration. In an increasingly competitive Tunisian banking market and an ever more demanding regulatory environment, UIB had to redefine its data foundations — with one certainty: AI could not simply be layered on top of existing systems.

‍

"Today, we can't do without AI. But if we want to work with it, we need to properly structure the process."

‍

UIB opted for an atypical approach for an AI project: starting with the technical foundations. Alteryx established itself as the go-to platform for combining data manipulation, modeling, and automation — accessible to both IT and business teams. Prime Analytics was chosen as a technical and functional partner to accelerate deployment.

‍

"With Prime Analytics and Alteryx, we chose to work together to operationalize several value-generating use cases."

‍

Three shared observations shaped the approach, as noted by William Rank, Director Channel Sales at Alteryx, drawing on a recent BCG study:

‍

"Only 20% of analytical models incorporate contextual data. However, AI is only effective if it understands the context. This is a real challenge for all companies, regardless of the sector."

‍

The obstacles are structural: data siloed in information systems, fragmented across business departments, and often in heterogeneous formats (tables, files, PDFs). The solution lies in a solid technical foundation, capable of facilitating collaboration between business and IT.

‍

With Prime Analytics and on Alteryx, UIB deployed a Smart Data Hub — a unified analytical foundation that centralizes all data from multiple systems in one place.

‍

The Smart Data Hub covers five essential functions:

• Accessibility to data sources with access control

• Data blending between internal data, Excel files, and external sources

• Governance over uses and transformations

• Automation of processing and orchestration

• Connection to AI models (LLM, proprietary models, cloud or on-premise models)

‍

"We first built the foundation with Alteryx workflows. And then, we implemented many use cases through what we call an automation factory."

‍

At the core of the foundation: Alteryx workflows for multi-source data collection (structured and unstructured), automated quality controls, data cleansing, and an operational governance discipline tailored to the banking context.

‍

Building on this foundation, the team deployed an automation factory: an industrial mode for producing business use cases, based on shared rules and a unified technical stack.

‍

Among the first use cases deployed:

‍

• Behavioral Customer Segmentation — unsupervised AI clustering on the Retail portfolio (K-Means)

• Corporate Internal Rating System — a probability of default model, complementary to existing SG models, adapted to the specificities of Tunisian regulations

• Internal IFRS Rating — extension of the rating system to IFRS requirements, with auditability and continuous monitoring

• Model Monitoring — Alteryx tools to track scores, detect deviations, and ensure the absence of AI hallucinations

‍

"Last year, we completed five use cases together. This year, we're aiming for a dozen with Prime Analytics, leveraging Alteryx. We are particularly pleased with this partnership — together, we are building great things for the bank."

‍

Mohamed Bessa also emphasized a crucial point: model evaluation must be both statistical AND business-oriented. A score, however impressive on paper, must also be defensible to regulators and understandable by operational teams.

‍