Case study · Generative AI · Under NDA

Designing an internal AI chatbot for a US banking client

Designing an internal AI chatbot for a large enterprise

Designing an internal AI chatbot for a large enterprise

An internal generative AI platform for about 9,000 employees at a US banking client.

Role

UX owner, end to end

Context

US banking client, internal tool

Users

About 9,000 employees

Confidentiality

Under NDA

What I can share

This project is confidential, so I can’t share the company name, real data or numbers. The screens on this page are sample designs with placeholder content, and the page focuses on the problems I solved and the value they created. I’m happy to walk through the process and decisions in conversation.

Overview

I helped a US banking client design and deliver an internal generative AI platform that helps about 9,000 employees securely explore, summarize, draft and analyze information using approved AI. It’s designed to improve productivity while supporting enterprise standards for data protection, responsible AI and internal knowledge access.

The result

The result

The result

The assistant improved employees’ everyday productivity. The exact figures are confidential.

Three problems, three design responses

Three problems, three design responses

Three problems, three design responses

Each challenge below shaped a part of the product.

Make AI useful for daily work

Make AI useful for daily work

Challenge

Generic AI tools weren’t tailored to employees’ daily work.

Value

Faster everyday work through prompt shortcuts, a prompt library and document generation.

Make AI approachable

Make AI approachable

Challenge

AI felt complex and intimidating, with guardrails and security rules that created confusion and frustration.

Value

An approachable, friendly personality, plus clear communication and disclaimers so people understood the boundaries and felt confident.

Give people one place for support

Give people one place for support

Challenge

There was no central place for communication, help and resources.

Value

A unified hub with guidelines, release notes, inspiration and everything users need for support.