BNP Paribas is taking a measured step deeper into artificial intelligence, introducing a new internal tool designed to support one of the most time-intensive parts of investment banking: preparing client pitches.
According to a report by Financial News, the bank has rolled out a platform called IB Portal, which uses AI to help bankers assemble pitch materials more efficiently. Rather than starting from scratch each time, teams can use the tool to search through BNP Paribas’s existing library of pitch decks, market analysis, and deal precedents to find relevant content faster.
Pitch preparation is a core task across investment banking, often carried out under tight deadlines. Bankers typically pull together market data, transaction history, and tailored messaging for each client, much of which already exists elsewhere in the firm. Yet that material is frequently recreated, leading to duplicated effort across teams and regions.
IB Portal aims to reduce that repetition. The system uses what the bank describes as “smart prompts” to surface relevant slides, charts, and analysis from past work, allowing bankers to reuse and adapt approved materials instead of rebuilding them.
George Holst, head of the corporate clients group at BNP Paribas, described the tool as an AI-powered search engine tailored for investment banking. He said it can shorten research and preparation time by days, giving teams more space to focus on strategic thinking and client judgement rather than document hunting.
Importantly, the tool is designed to fit within the constraints of real banking workflows. Pitch decks are not generic presentations; they include internal viewpoints, client-specific details, and regulatory considerations. As a result, IB Portal emphasizes structure and control over creativity. Access is governed by region and business line, and any content surfaced by the system must still be reviewed by humans before it is shared externally.
Traceability is another key feature. Bankers can see where information originates, helping reduce the risk of errors or unintended disclosures. Without those safeguards, the use of AI in sensitive environments like investment banking could quickly create compliance and reputational issues.
Part of a broader internal AI strategy
IB Portal sits within a wider push by BNP Paribas to develop AI tools on its own infrastructure. In June 2025, the bank announced an internal “LLM as a Service” platform that gives business units access to large language models hosted in BNP Paribas data centres, supported by dedicated GPU capacity.
Run by internal IT teams, the platform supports a mix of open-source models and systems from Mistral AI, with plans to introduce models trained on the bank’s own data. Intended uses include internal assistants, document drafting, and information retrieval, all within a controlled environment.
BNP Paribas is not alone in this approach. Other global banks are building similar in-house systems. JPMorganChase has expanded access to its internal “LLM Suite,” while Goldman Sachs has invested heavily in AI engineering and launched its own “GS AI Assistant.” UBS has discussed an internal M&A co-pilot, and specialist AI tools such as Rogo are gaining adoption at firms including Nomura and Moelis.
For BNP Paribas, the key question is whether IB Portal becomes part of everyday banking work rather than a limited trial. The upside is clear: less time spent searching for information, fewer duplicated pitch decks, and better use of institutional knowledge. The risks are equally well understood, from inaccurate outputs to unclear sourcing or exposure of sensitive data.
Banks that have seen the most stable results tend to keep AI tightly constrained, grounding outputs in approved internal content, applying strict access controls, tracking usage, and requiring human sign-off before anything reaches a client.
