← Cases
GermanyFinance and insurance2024 – ongoing

Deutsche Bank

DB Lumina research agent

Deutsche Bank's DB Lumina reaches about 5,000 research staff; analysts report 30 to 45 minutes saved on earnings-note templates

WorkingVendor's case studyVendor ↗

Problem

Research analysts spent time on repeated document work, and the bank needed any AI tool to meet data privacy rules for a regulated lender.

Approach

Deutsche Bank built DB Lumina with Google Cloud Consulting on Gemini models. It has a chat interface, prompt templates for documents, and a retrieval feature that cites its sources. It went live in September 2024.

Results

~5,000Users across Deutsche Bank Research; target of 10,000+ by end of 2025Vendor's case study[1]
30-45 minTime analysts say they save preparing an earnings note template (up to 2 hours on research reports and roadshow updates)Vendor's case study[1]

Takeaway

The bank started with a pilot, then rebuilt the tool around retrieval with citations. Analysts estimated the time savings themselves.

Categories

Innovation type
Process innovation
AI technology
Generative AI
Evidence
Vendor's case study

Sources

  1. Deutsche Bank delivers AI-powered financial research with DB Lumina
    Google Cloud Blog24 Sep 2025Vendor publication

Source summaries

[1]Google Cloud Blog24 Sep 2025

A Deutsche Bank AI engineering head and a Google author described how DB Lumina was built. They gave the user count and the time savings analysts reported.

Full details are in the original sources.

Added 25 Oct 2026 · Updated 25 Oct 2026
Open case page ↗