Responsive experience designs delivered as HTML for direct integration into production code

Secure email becomes responsive – With the help of iterative designs and UX discovery work my squad was able to re-tool this product. We were able to simplify this experience and also reduce the volume of emails sent into to us.

Lead designer
Chat-bot integration
Data analysis
Design systems
UX Research planning
Machine learning

Project background

Project background

Fidelity customers have all sorts of questions, they need to deliver documents and personal information securely. We only want customers to use secure email for sending in personal information and documents.

We integrated with a machine learning API and a virtual chat-bot to clarify the customers question. When we provide customers with a matching response to their intent, we can initiate a conversation enticing them to find their answer.

If this is successful, we expect that solution will deflect about 10% of customers away from sending question-based emails saving time for the customer and save money with the email response team.

Access these pages live

If you are Fidelity customer, go to Customer Support > Contact Us > Send secure email

If your not a customer, I can share in person.

Problem

Responding to emails is very costly for the firm. We only want certain types of emails, so we are encouraging customers to self-serve and find answers another way.

Our problem is that we receive too many general question-based emails that could be answered by the Chatbot or content on the site. The cost of responding to emails keeps increasing as our customer base increases. The support team is stressed and it is simply not a sustainable long term model.

Goal

In MVP2 we needed to reduce quantity of emails overall, reduce all general question-based emails that can be answered well by the chatbot and or the website.

(MVP 1 was focused on a new form UX and technical service.)

Early results

We did see a shift in the type of emails with less money movement transactions and trading requests. Personal data sharing and document uploads related to specific activity did start to increase. The chat-bot was only able to catch and deflect a small number of emails.

Detailed production designs and scope

Default view small breakpoint

Users arrive here from Contact us

Default desktop view


Artificial intelligence response starts

After users input a subject, we  start processing and clarify their intent with a “Did you mean” response.

Understanding customers intent

Users may converse with the Virtual Assistant and find the answer they are looking for.

Confirmation

When customers do not want to enage with the Virtual Assistant they can simply close the VA and easily submit an email with a simple, optimized form.