dbLumina — MVP 0→1

The Investment Banking division at Deutsche Bank operates in a hyper-competitive and information-rich environment. The ability to rapidly access, synthesize, and act upon vast quantities of financial data is paramount to maintaining a competitive edge, advising clients effectively, and identifying new opportunities.

Currently, professionals across divisions (analysts, associates, research staff) spend a significant amount of their time on manual, low-value tasks involving data gathering and processing.

From a product and design perspective, this meant balancing speed with accuracy, simplifying complex AI interactions in high-stakes financial workflows, in order to reduce cognitive load without removing user control or transparency.

dbLumina is a secure, dedicated GenAI analyst at their fingertips — a web-browser application that uses Google's Gemini large language model while supporting customisation, such as fine-tuned skills.

[How the MVP was scoped around a small set of core capabilities, with fine-tuned skills co-designed with business units.]

Turning generic AI into repeatable enterprise workflows.

Designing under governance and compliance constraints.

Designing trust in a probabilistic system.

Balancing simplicity with powerful capabilities.

Designing for multiple levels of AI maturity.

Defining interaction patterns for “agentic” experiences.

[What shipped in 2024, who adopted it, and how the fine-tuned skills validated the platform direction.]