An industry where projects are run under supervision
IT expertise in banking and insurance has one particular feature: the end user is not the only judge of the result. A credit institution, an insurance undertaking or an asset management company must be able to demonstrate to its supervisor that it controls its information systems, its providers and its data. An application that works is not enough. You need evidence that it was designed, tested and operated according to known rules.
This requirement changes the nature of the work. The risk, compliance and security functions are not end-of-project reviewers: they are stakeholders in the framing. A project that calls on them on the eve of go-live discovers too late that it is missing an analysis, a clause or an audit trail.
On top of this comes the weight of the existing estate. Account management, policy administration and claims management systems concentrate business rules accumulated over the years, often poorly documented. Anything new (an online customer journey, a scoring model, a cloud service) has to work alongside them.
The three texts that frame your projects
DORA. Regulation (EU) 2022/2554 on digital operational resilience for the financial sector has applied since 17 January 2025. It organises the management of information and communication technology risk, the handling and reporting of major incidents, resilience testing and the control of third-party providers, with a register of information on contractual arrangements to be kept up to date. Our glossary entry on the DORA regulation sets out its structure.
GDPR. Regulation (EU) 2016/679 applies to the data of customers, prospects and employees. Two provisions weigh on the design of financial systems: Article 22, which governs decisions based solely on automated processing, and Article 35, which requires an impact assessment when processing is likely to result in a high risk to individuals.
AI Act. Regulation (EU) 2024/1689 classifies AI systems by level of risk, with phased application. Its Annex III lists as high-risk the systems that evaluate the creditworthiness of natural persons or establish their credit score, excluding financial fraud detection, as well as those used for risk assessment and pricing in relation to natural persons in life and health insurance. The details are in our entry on the AI Act.
These three texts are not three separate compliance workstreams. They ask the same question from three angles: do you know what your system does, who is accountable for it, and can you prove it? A well-framed project organises its evidence once and reuses it.
Typical workstreams for an IT department
- Modernising the core system. Replacing or progressively breaking down a central application, while maintaining service and accounting consistency at every stage. This is the purpose of our application modernisation offering.
- Controlling the provider chain. Listing the services entrusted to external providers, identifying those that support a critical or important function, preparing the exit before it is needed. This work falls under cyber resilience as much as under procurement.
- Making management and reporting data reliable. The same data feeds management, risk measurement and reporting to the supervisor. Its definition, its owner and its traceability fall under data governance.
- Putting AI into production without losing control. Inventorying AI systems, classifying them, documenting models and organising human oversight: this is the role of AI governance.
What this changes in the way a project is run
Three habits set a financial project apart from an ordinary digital project.
Evidence is a deliverable. Architecture file, risk analysis, test results, committee decisions: these documents are produced as the project progresses, not in a catch-up exercise. Our consultants help put them on the schedule with an owner and a due date, like any other work package.
The calendar is not free. Accounting closes, reporting deadlines, change freeze periods: deployment windows are rare and known in advance. The schedule is built backwards from these windows, not from the hoped-for end of development.
Dependencies dominate. The same customer master data or the same calculation engine serves several projects at once. Without an overall view, two teams modify the same component in opposite directions. Major transformations are therefore managed as programmes, with programme management that arbitrates priorities and shared resources.
IT expertise in banking and insurance: how we work
Management Square’s roots are in project, programme and portfolio management advisory. This background matters in an industry where the difficulty lies less in choosing a technology than in getting it into production within a constrained framework. We work across our five domains: artificial intelligence, data, cloud and platforms, software engineering, transformation management.
Our approach follows four stages. Frame: clarify the business objective, the regulatory scope and the functions to involve. Prove: check on a reduced scope that the solution keeps its promises, including in terms of security and traceability. Industrialise: scale up with the expected controls, documentation and operations. Operate: monitor the service over time, with operations that rely on your teams, to whom our consultants pass on their practices.
Interpreting the texts is a matter for your legal and compliance functions. Our consultants help your teams translate their requirements into architecture choices, a project plan and usable evidence. To discuss your challenge, get in touch with us.