Wissem Souissi

The through-line

The recurring argument is that artificial intelligence in investment management earns its place only by disappearing into the work people already do. Dasseti's AI effort began "around 18 or 24 months ago" and was later "formalize[d] ... into a long-term sustainable strategy supported by uh robust infrastructure, thoughtful features and entitive UX and of course seamless integration" [1]. The timing is presented as deliberate rather than reactive: "the technologies available today are much stronger than 2 or 3 years ago", the industry is "more ready uh for such kind of changes", and clients now "have more clear ideas on how they would like AI to leverage" in daily use [1]. That readiness on all three fronts, technology, market and buyer, is what turns years of scattered development into a single embedded assistant. It sits alongside the underlying business preoccupation of helping LPs and GPs in private equity make better use of their data [2].

On building AI into the workflow rather than on top of it

The design principle behind Sidekick is that users should not have to go anywhere new to get value from it. "Rather than being a standalone kind of application, uh, Sidekick operates directly within our different solutions uh, collect, engage, and and and DSG's platform uh, Harvest", so that "users will never have to leave their normal workflow uh, across the different kind of modules" [1]. The reach spans due diligence, investor relations, compliance and data management, and the phrasing is consistent: the assistant is integrated into real workflows rather than layered on top of them [1].

On structured outputs and institutional knowledge

The more distinctive claim is about what the AI leaves behind. Sidekick "creates uh structured outputs that become valuable assets over time", which supports "the kind of building the institutional knowledge rather than just providing answers" [1]. That reframes the value of an assistant away from the moment of the query and towards a compounding record that a firm accumulates, which is consistent with the wider concern of improving how private equity investors use their data [2].

On measuring AI by value, not by how impressive it looks

There is a firm test applied to every feature. "AI features are measured by value added not how flashy they look", and "if an AI solution doesn't improve how investors or fund managers um interact with data or streamline compliance while maintaining full trustability. Of course, it's a no-go for us" [1]. Trustability sits inside the acceptance criterion rather than beside it. The method for enforcing this is empirical: "we test with real user workflows", and the stated discipline is that "we don't build features for the sake of ticking a box. We solve specific problems our clients face daily" [1]. Success is defined accordingly, "not by how impressive our AI appears in isolation uh but how effectively it helps clients make better investment decisions, build stronger relationships and manage risk uh more efficiently" [1].

On clients as co-developers

The reported effect on client relationships goes beyond productivity. Clients "see sidekick as a strategic differentiator uh not just an efficiency uh tool", and they are "coming to us with their own ideas ... for you know uh how AI can support their unique processes", which is treated as "a co-developed kind of journey" [1]. The AI is described as changing the conversation and the relationship clients have with their data [1].

Takeaways

• AI should be embedded in existing modules so "users will never have to leave their normal workflow", not delivered as a separate application [1]
• The value of an assistant lies in "structured outputs that become valuable assets over time", building institutional knowledge rather than answering one-off questions [1]
• Any AI feature that fails to improve data interaction or streamline compliance "while maintaining full trustability" is rejected [1]
• Features are validated by testing against real user workflows, not by ticking boxes [1]
• Success is measured by better investment decisions, stronger relationships and more efficient risk management, not by how impressive the AI looks in isolation [1]
• Clients now treat the AI as a strategic differentiator and bring their own use cases, making development a joint effort [1]
• The formalisation of an AI strategy waited on three things maturing together: the technology, the investment management industry's readiness, and clients' clarity about what they wanted [1]