Article first published in 𝘍𝘪𝘯𝘢𝘯𝘻 𝘶𝘯𝘥 𝘞𝘪𝘳𝘵𝘴𝘤𝘩𝘢𝘧𝘵, February 7th 2025
The certificate industry is also being revolutionized by artificial intelligence.
The Swiss certificate industry has always been a pioneer in automation. AI promises to push these boundaries even further, starting from an area in which it has already achieved its first major successes: software development. Companies that use tools such as GitHub Copilot, ChatGPT, or Claude report productivity gains of over 30% through faster code generation, improved vulnerability detection, and automated testing. But that’s just the beginning.
In the field of document creation, AI not only accelerates traditional templating, but also enables automation for documents that were previously created manually. These include less frequently used issuance programs, as well as marketing materials where automation costs have been too high in the past, or markets such as the U.S., where the influence of law firms has made automation more difficult.
Many areas of application
Another area of application is the further processing of documents that have already been created, for example the adaptation of PRIIP KIDs to the local requirements of national authorities, which go beyond the European requirements. AI is also opening up new opportunities in the complex area of ESG documents. Not to mention non-retail-oriented documents such as OTC term sheets.
AI will also have a decisive influence on data mapping, which traditionally takes up a large proportion of banks’ annual IT budgets. Whether communicating with regulators, exchanges, or third-party platforms, generative AI can significantly speed up mapping processes by analyzing the underlying documentation, capturing the relevant schemas, and generating the appropriate rules. Especially from the point of view of the recently enacted DORA Regulation, AI can play an important role in risk management. In the field of data reconciliation and life cycle management, AI’s ability to analyze and detect inconsistencies across disparate data sets, including documents such as final terms, term sheets, and paying agent emails, helps banks avoid costly mistakes and improve customer service.
Perhaps one of the most exciting applications is in setting up new investment products, which is typically done manually. Here, AI can recognize product types and data by processing customer emails and analyzing call recordings, automatically set them up in an issuer’s trading systems, thus significantly reducing time-to-market and minimizing human error.
Apart from the backend optimizations mentioned above, AI is also revolutionizing the way bank employees will deal with IT in the future. Instead of having to master several applications, employees will be able to use AI-controlled chatbots in the future to directly access the functions that are relevant to them. This not only increases employee productivity, but also expands their radius of action, as the occasional user can work just as effectively as the daily user.
A wide range of applications are also opening up in the customer area. Who doesn’t know them – the bots that dial into virtual meetings and record the conversations. In the area of certificates, however, the possibilities go even further. During sales calls, AI assistants can ensure compliance with regulatory requirements regarding the appropriateness and suitability of a product, while translating customers’ market views and desires into product ideas in real time. Even futuristic-looking ideas such as sales talks autonomously led by AI are already feasible from a technical point of view.
But not only sales, but also customer service benefits. While banks generally handle customer complaints better than other industries, AI can further improve the customer experience by responding to queries faster.
Future in the industry
So what’s holding back the widespread adoption of AI? The primary hurdles are institutional in nature. Banks need time to develop internal AI policies and find answers to fundamental questions, such as whether to base their AI strategy on commercial models or open-source alternatives. Standards such as ISO 42001 for AI management systems should speed up this process. There are also many interesting innovations in the hardware area, such as Apple’s M4 Max chips and Nvidia’s recently announced DIGITS project, which make it possible to run even powerful models in a bank with relatively little effort.
The problem of hallucinations should also be manageable through the use of modern technologies. These include techniques such as retrieval-augmented generation (RAG), in which AI answers are anchored in verified data, the use of deterministic tests in relation to data generated with AI, but also human validation steps in particularly critical processes.
So the question is not so much whether AI has applications in the certificate sector, but how quickly banks will implement them.
