Questions & Answers

AI can concretely improve relations with investors and IR teams, provided it is used as a lever for reliability and consistency, and not as a tool for automated message production.
IR teams face a growing requirement to respond faster, provide accurate, consistent and contextualized information, while adapting to very different investor profiles. In this context, AI can play a structuring role.
In concrete terms, it can be used to prepare summaries from internal reports, to reformulate content according to the recipient's level of expertise, to quickly find information in the history of exchanges or in a data room, and to improve the overall consistency of documents sent. It can also assist in the production of standardized responses (FAQs, standard emails), while maintaining a high level of editorial quality.
But the real contribution of AI does not lie in speed of execution. It lies in the ability to align messages. High-quality IR communication relies on single, reliable data shared between teams. If AI is plugged into fragmented or poorly governed sources, it amplifies inconsistencies instead of correcting them.
The challenge is therefore to anchor AI in a controlled data chain: same figures between BI, reporting and investor communications, traceability of sources, and systematic editorial control before sending. In this context, AI becomes a powerful support tool for structuring, harmonizing and securing communication.
The right balance consists in using AI to prepare and make content reliable, while leaving IR teams responsible for tone, context and relationship. It is this combination that improves both operational efficiency and investor confidence.

Effectively training fund teams in AI is not about imparting theoretical knowledge, but about transforming concrete work practices.
Relevant training always starts with the real situations encountered by teams. Investment professionals don't need a general discourse on AI, but an operational understanding: what the tool actually enables, its limitations, and the conditions under which it can be used without degrading the rigor of processes.
This means segmenting approaches. The needs of a partner, an analyst, an IR, compliance, middle office or ESG team are profoundly different. Effective training is therefore based on a common foundation (principles, risks, best practices), supplemented by targeted use cases: analyzing an investment memo, summarizing an Information Memorandum, exploring a data room, preparing for a committee, sector screening or managing a complex exchange with an LP.
The key is immediate applicability. Each module must enable action to be taken the very next day, with visible and measurable gains. This is what transforms acculturation into real adoption.
But training cannot be seen as a one-off event. Models evolve, tools change, uses become clearer and risks shift. An effective approach requires a long-term approach: initial awareness-raising, practical workshops by business line, feedback from peers, and ongoing support to adjust practices.
Finally, a point that is often underestimated: training in AI also means training in discernment. Knowing when to use the tool, when to be wary of it, and how to control its results is just as important as knowing how to use it.
The right system therefore combines teaching, practice and iteration. It is this logic that enables AI to be firmly anchored in a fund's processes, without falling into a theoretical approach disconnected from the field.

A classic SaaS executes pre-programmed actions: if X, then Y. An AI agent, on the other hand, receives an objective, understands the context, plans its steps and executes a sequence of actions, adapting to the data it encounters. This autonomy is both its interest and its risk.
In a fund, the relevant use cases for an agent are well-defined: processing supplier emails, enriching a deal sheet, automatically preparing a committee pack from a data room, monitoring signals on target companies. Ill-framed cases (making an investment decision, sending a communication to an LP without proofreading) are not cases for agents, they are cases for assisted humans.
A good agent has four characteristics: a clear perimeter, human supervision at every sensitive stage, complete traceability, and reversibility in the event of unexpected behavior. Bodic develops its agents according to these principles as part of its customized developments.

Hallucinations are the main risk when using AI agents in a financial context. They are not the result of random errors, but of the structural behavior of models in the absence of reliable or sufficiently constrained information.
Three levels of control can greatly reduce this risk.
The first is strict anchoring on a certified source of truth. The agent must not rely on general knowledge or implicit data, but only on a controlled internal repository. Each answer must be associated with an identifiable, accessible and verifiable source. An unsourced answer must be considered invalid by default.
The second level is the restriction of the functional and informational perimeter. An agent must intervene in a precise domain, with a limited and controlled set of data. The wider the scope, the greater the risk of approximate interpretation. In practice, an agent who specializes in a subset of financial data, such as a shareholding's KPIs or Middle Office flows, is significantly more reliable than a generalist agent.
The third level is the implementation of systematic controls on outputs. All numerical information intended for external use, in particular for investors, must be validated by a human being. This validation must be based on complete traceability: initial request, sources used, transformations applied and response generated.
Beyond these principles, technical architecture plays a decisive role. A structured approach consists in isolating the AI layer from raw data, using a centralized repository, then exposing only validated data to agents. This allows precise control over what the agent can see and use.
Finally, it's important to treat hallucinations as a governance issue, not just a technical problem. This means defining rules of use, levels of responsibility and supervisory mechanisms adapted to the financial stakes involved.
A reliable agent is not one that "responds well", but one whose every response can be explained, traced and verified.