Fidelity Investments
· FintechGenAI Workflow Automation
Agent-assisted workflows for enterprise operations teams
- Faster task completion
- 25%
- Users on the platform
- 200K+
Context
Full case study in progress. Cloud platform operations teams at Fidelity spend large parts of their day on repetitive, multi-system workflows. This project designed GenAI-driven automation for those internal workflows, exploring where assistance genuinely helps and where it erodes trust.
Design tension
The hardest problem was deciding what NOT to automate.
- Automated the repetitive, multi-system steps where errors were mechanical, not judgment-based.
- Kept humans in control of consequential decisions: approvals, exceptions, anything with blast radius.
- Designed the handoff points so operators always knew what the system did on their behalf.
Outcome
The shipped automation cut task completion time by 25%. A full write-up covering the research, the human-in-the-loop design decisions, and what I would do differently is coming soon.
Fidelity Investments maintains a strict policy around sharing internal work publicly. Details in this case study are generalized in compliance with signed non-disclosure agreements.
Seen work you'd want on your team?
I'm interviewing now. The fastest way to reach me is a 30-minute intro call, or drop me a line and I'll reply within a day.