Eneco: a cockpit for intraday traders, with humans at the wheel

The problem

Intraday renewables traders work under constant time pressure, across multiple screens at once, in a market that changes by the quarter hour. Their way of working had grown over years: trusted column layouts, color codes with fixed meaning and personal tools next to the official systems. Meanwhile automation was advancing, while trust in it was limited. In the words of one trader: “De autotrader heeft geen marktinzicht of intuitie.” (The autotrader has no market insight or intuition.)

The assignment: a new cockpit that can be scanned in seconds, and an interface that gives traders real say over when automated trading may intervene. Without breaking the conventions they have relied on for years.

My role

As enterprise UX designer and researcher I was involved from discovery onward: from user research and service mapping to prototyping, validation and decisions around UI libraries and technology.

The process

Taking years of habits seriously

I interviewed and observed traders to understand what actually carries their way of working. Many conventions turned out to be functional: the fixed position of tables and charts and the meaning of colors are not taste but a shared mental model, built over years of working together under pressure. The design had to innovate within that, not against it.

Information density versus scannability

The reflex with expert users is to show more data. The research pointed the other way. A dashboard that must be scanned in seconds calls for reduction: fewer columns, smarter defaults and detail that only appears when asked for. I put those choices in front of the whole team in a group session, which ended with concrete, prioritized design decisions instead of an endless feedback loop.

Designing trust in automation

For the Autotrader interface I facilitated design sessions and captured the requirements as job stories. The core: automation is accepted when the trader can see what the system is doing, can understand why, and can intervene at any moment. Control has to feel accessible, not like an emergency brake.

Working prototypes instead of static sketches

With AI tooling I built interactive prototypes in a short time, so traders could respond early to a working version with realistic interaction instead of to a picture. For users who think in seconds, that is the difference between polite nodding and real feedback.

The solution

A position dashboard that brings together live position, forecasts and autotrader status across sub-portfolios and bidding zones, designed for scannability under pressure. And an Autotrader layer that makes the collaboration between trader and automation explicit: when the system acts, when the human does, and how you switch.

Results

The group session produced supported, prioritized design decisions and broke the feedback loop open towards a first release. The AI-accelerated prototyping approach became the standard way of working within the project to validate requirements early with the people who have to work with them daily.

What I take away

With expert users, resistance to automation is rarely stubbornness: it is knowledge of the exceptions the system does not know. Whoever takes that knowledge seriously in the design, from color code to override, builds automation that actually gets used.