We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·May 6·5 min read
Singular Bank helps bankers move fast with ChatGPT and Codex
TL;DR
Bankers spent 60–90 minutes daily on routine tasks like meeting preparation, portfolio analysis, and follow-up work. Singular Bank built Singularity, an internal assistant combining ChatGPT (a conversational AI model) and Codex (a code-generation model) to automate these workflows.
✦ Why It Matters
Engineers can apply multi-model AI combinations to automate repetitive knowledge work and measure productivity impact quantitatively.
Key Takeaways
- Combine conversational AI and code generation models for complementary automation capabilities.
- Measure productivity gains in minutes saved on specific, repetitive tasks.
- Design AI assistants to integrate into existing workflows rather than replace them.
- Apply large language models to knowledge-worker roles in regulated industries like banking.
Full Summary
Singular Bank identified that bankers were spending significant daily time on repetitive administrative work including meeting preparation, portfolio analysis, and client follow-up—tasks that didn't require specialized financial judgment. To address this, the bank developed Singularity, an internal assistant leveraging two OpenAI models: ChatGPT for natural language understanding and task automation, and Codex for generating code snippets and automating data processing workflows.
The assistant integrates into bankers' existing workflows, handling routine information gathering and document preparation. Results showed concrete time savings of 60–90 minutes per banker daily, freeing capacity for higher-value client interaction and strategic analysis.
This demonstrates how large language models can be applied to knowledge-worker productivity in regulated industries.