CONTEXT / Key Terms
AMEX is investing heavily in integrating AI solutions into their internal software used for servicing card members.
Like many enterprises, AMEX is looking for ways to leverage AI to stay competitive and improve the quality of their services. In 2024, we've helped AMEX create four MVP solutions that integrate AI into their existing software.
Interactive Voice Response (IVR)
IVR is the system card members talk to on the phone that gathers basic info, before being directed to the right department at AMEX
Intent
The Intent is what the caller wants to do, or why they're calling. For example, to "Book a flight", "Cancel a Dining Reservation", "Check Account Balance", etc.
Just-in-time (JIT) UI
JIT UI delivers information, options, or features to the user precisely when they are needed, rather than presenting everything upfront.
Card Member (CM)
A CM is a person that has an AMEX card.
AMEX's goal
Leverage AI to provide travel agents with rich, just-in-time data, allowing them to offer a more personalized booking experience for their card members.
AMEX has the competitive advantage of a vast and rich database of information on their card members' travel habits. Their goal is to use AI to provide their travel agents with actionable, just-in-time data to foster a more personalized booking experience for their card members.
PROBLEM
AMEX travel agents are ignoring the card member dashboard, which includes the insights needed to provide a personalized booking experience.
Despite having the competitive advantage of a rich data set, travel agents are not leveraging it. The main reason is that the dashboard is "messy" but not comprehensive. In other words, the needed data points are not localized or quickly accessible, and the existing data is ignored due to a poor UI.
Old Card Member Dashboard
Travel Agents Thoughts:
“Doesn’t have enough data to be useful on it’s own."
“I ignore the dashboard. Our calls are too hectic.”
"It's a bit messy. I prefer to have my own notes."
"Spending Trends would be useful, but the design makes it complicated."
HYPOTHESIZED SOLUTION
An AI just-in-time card member dashboard that displays information based on the card members' intent captured in the IVR.
For example, if the card member called to book a flight, the dashboard would display preferences like their "Preferred Airline" for a personalized experience. If they called to cancel a flight, it would show their upcoming flights.
Discovery
Creating a tagged database with all relevant data points would help us determine which information to show based on user intent.
This was a critical first step not only for us designers but for for the engineering team in the handoff phase.
Discovery Results
A completely AI driven dashboard would not solve user painpoints or achieve AMEX's goals of creating a more personlized seemless cexperience becuase:
💡
Travel Agents need the flexibility to book multiple event types at once.
A card member might call to book a car service and a flight at the same time, but the IVR only captures one intent. So showing data based on one intent would have a negative impact on the travel agents ability to provide a personlized experience.
💡
The IVR doesn’t always capture the correct intent
Most of us have experience an IVR system not understand us correctly. Showing data for the wrong intent would result in a negative impact on the travel agents ability to provide a personlized experience.
Adjusted Solution
A dashboard that leverages progressive disclosure by segmenting data by event type.
This organization of data allows for flexibility and makes it easy to find the most relevant data points for each event type. The dashboard still leverages speech-to-text by navigating the travel agent to the correct event type based on the card member's intent.
MY IMPACT/ROLE
I was the Design lead responsible for facilitating cross functional collaboration between Product Designers, Product Managers, Engineers.
Balanced the needs, goals, and restrictions of AMEX, AMEX product owners, Engineers , travel agents, and card members.
Led a team of 2 designers and 1 project coordinator
Co-lead the UI design process
