Eric Bosse’s inspiring speech on the next generation of trading desks was a powerful opener at CMX 2025. What he showed us was a blueprint that combines technology, people and process in ways most desks still treat as science fiction. Here are the seven ideas that matter.
1. Start by imagining your ideal desk
Even if you’re not building from scratch, it pays to think like you are. Bosse challenges leaders to sketch out the team, tech and workflows they would build if nothing was in place yet. This helps expose inefficiencies, clunky systems and legacy decisions that are holding performance back. It’s a mindset shift from maintenance to design.
2. AI and humans are becoming colleagues, not competitors
Bosse’s team has already embedded AI into day-to-day trading decisions, including live algo recommendations. Over time, the human traders improved their performance by learning from the machine’s strengths. The two sides became more effective together. The lesson is simple: AI isn’t replacing traders, it’s supporting them.um data fields, such as T+0 enrichment criteria, for reliable automation. Additionally, Fixed Income tokenisation or a central limit order book (CLOB) were discussed as long-term solutions to support operational flow and efficiency.
3. You no longer need traders who can code – you need traders who can think
For years, coding was a prized skill on the desk. That’s now shifting. Bosse argues that it’s more useful to hire people who can ask the right questions of the data, even if they can’t write the code to extract it. With copilots and no-code tools taking over simple scripts, human insight becomes the real differentiator.
4. Proprietary tech is no longer a competitive edge
Building your own EMS used to be a mark of distinction. Today, it often means unnecessary risk and complexity. Bosse’s advice is to avoid owning infrastructure wherever possible. Off-the-shelf tools can now be configured to suit most strategies, so focus on integration and smart configuration, not reinvention.
5. AI can be a surprisingly good coach
In an internal experiment, Bosse’s team compared trader decisions with machine-generated recommendations. Over time, the traders adopted better habits, made fewer errors and didn’t revert when the AI was removed. AI helped shape better human decisions, not just faster ones.
6. TCA is essential if you want to prove trading alpha
You can’t improve what you can’t measure. Bosse is clear that TCA isn’t about oversight – it’s about clarity. As more execution becomes automated, traders need to show where they’re adding value. Without data, that story falls apart.
7. Creativity matters more than ever
Quant skills and market knowledge still count, but Bosse argues that creativity is the trait that will set the best desks apart. With the pace of change accelerating, the ability to rethink problems, test new tools and adapt quickly is becoming essential. Machines may be learning fast, but they still can’t connect dots the way humans can.
