As technology evolves, it offers many opportunities to streamline, secure and expand the tools we provide staff and the services we can offer clients. Remaining at the cutting edge of technology development and uptake means trialling and the rapid adoption of the most promising features. This ensures our group and our clients benefit from both a first-starter advantage and the ongoing efficiencies and security that technologies bring. In this article we shine the spotlight on CACEIS’ ‘AI Factory’ that has been a feature of our Digital division since 2024.
The AI Factory was conceived early last year in order to ensure in-house and client-focussed needs around Artificial Intelligence were identified, understood and met in a systematic and efficient manner. Our AI Factory has identified in excess of 100 use cases from its comprehensive surveys of CACEIS’ various business lines and individual staff members. After assessing the requirements across three key metrics – value, complexity and ease of integration, 4 specific groupings were identified and each use-case was siloed into one of the 4 categorisations known as ‘tech-enablers’. The tech-enabler categorisations are as follows:
1) Document reader with data extraction, to retrieve key information from both structured and unstructured documentation.
2) Static content generation, to provide standard responses to the most frequently asked questions for use by AI-powered chat interface tools and questionnaires.
3) An AI-powered content generator, able to link in to a wide range of databases to provide reports, data analytics and other content on a static or dashboard-style basis.
4) Complex/sensitive modules, for higher value or higher risk areas such as automatic reconciliations or multi-faceted fraud detection.
Using the AI Factory’s systematic assessment procedures, CACEIS has already put 4 fully-tested Artificial Intelligence technologies into production – and the results have been particularly positive:
PickAI, AI software developed in-house by CACEIS, reads and extracts data from invoices, which after validation or correction following our usual 4-eyes principle, can then be injected directly into the payments system.
Betsie, another in-house developed AI tool, enables our teams to efficiently draft business proposals in response to the many questions involved in Requests for Proposal (RfP) processes or Due Diligence questionnaires. To respond to questions posed, Betsie searches its vast database of validated documents from previous business proposals, requiring only a rapid validation to ensure the information is still correct and up to date. Betsie also allows for a full RfP/questionnaire to be uploaded and it will assess and provide responses for each question systematically.
Froggy is a chatbot developed to provide rapid responses to queries in our client-facing web portal, Olis. This in-house developed technology is designed to meet the needs of both internal users from CACEIS’ client support desks and of clients who are able to directly pose questions and receive clear responses to their queries. To provide the responses, Froggy is trained on our FAQ database, our vast library of procedural and operational documentation, and finally the training manuals and constant feedback from our staffed helpdesk.
ProspAIr is a fund prospectus reader, designed to extract key information from any type of investment fund prospectus in order to rapidly digitise the contents and provide summaries of investment policies, investment limits, and more to help set up and verify the fund for depositary control purposes and much more. ProspAIr can recognise and extract over 100 data points and after a standard 4-eyes verification, the information can be rapidly injected into our systems.
Sabine Iacono, Group Head of Digital, commented : “Throughout 2024, our AI factory has shown itself capable of the successful implementation of Artificial Intelligence-based technologies, both classic and generative to enhance, speed up and secure services for our own staff and our clients. We are delighted that so many of the projects implemented have been based on in-house developed AI, as this provides essential isolation from a potential bust of a technology partner in today’s fast-growing but potentially unstable AI start-up environment. Furthermore, by maintaining the all-important 4-eyes principle, we ensure our staff are assisted by AI rather than replaced by AI, which would represent a security and accuracy risk to our services.”
2025 will continue to see CACEIS identify and assess new use cases in areas such as automatic reconciliation, knowledge management, and prospectus drafting for clients.
Our digital division is based on the principle of ‘Build, Operate, Transfer’ in order to enable our internal clients to take ownership of their technology needs once implemented. This is key to the digital division being able to continue to innovate and focus on new value-added areas and increase the scale and scope of our support. Although many people see AI as an industry buzzword, these concrete and practical implementations are proof that in 2025 and beyond we are in an excellent position to leverage technology to support our staff and our clients.
