Building ✨Assist

Showing the design process behind our new customer response toolbar

Product Design
Design Systems
AI
Role

I was the lead product and visual designer on this project.

Team

3 engineers, 2 PM's and I

3 engineers, my PM and I

Timeline

Jan 2023 - Present (two phases)

Background

Our team explored implementing several of AI's applications in our products, starting with an MVP suite offering LLM-based comment analysis, chatbot and executive summaries.

Challenge

Due to a finite budget, our team was to compromise and hone in one specific application of AI that would be most lucrative to our customers in the shorter term.

Output

A set of basic customer response tools based around the ability to reply to a customer with unprecedented speed using AI.

Results

  • Reduced time to draft a customer response by 80%

  • Increased two major clients' satisfaction metrics by 15% and 6% respectively, during MVP stage

  • Successfully built a white-labelled AI design system, compatible with all our brands' CX products

Design stages

  1. Manual PDF reports showcasing customer insights

  1. Tested first online reporting MVP, complete with chatbot

  1. Pivot to customer response tool based on budget and user feedback

  1. Iteration to Assist product and launch

Pivoting focus to customer response

Qualitative insights from testing our MVP revealed clients were primarily keen on using the AI chatbot portion of our MVP to draft swift but detailed responses for their customers.

Iterations and launching ✨Assist

Initially, we assumed this feature would be a simple "Write with AI" button embedded in existing response fields throughout our apps.

Benchmarking our ideas against competitors and AI response tools in the wild revealed a longer list of requirements and constraints we needed to solve for with this feature.

After several major iterations, we reached a solution where the AI options were to be bottom-aligned within the text field. This orientation felt more subtle and not as forced. Also, placing the toolbar at the bottom of the text field does not obstruct users from writing a response on their own.

There were several constraints that were taken into consideration during this process:

  • adhering to existing style guidelines and design language with some wiggle room for all-new AI branded components

  • some of our user base is skeptical about incorporating AI into their workflow - how can we ease their doubts and concerns?

There were several constraints that we needed to be wary of during this project:

  • all the specific, existing use cases that the Assist toolbar needed to fit seamlessly into

  • adhering to existing style guidelines and design language with some wiggle room for all-new AI UI components


Launched version of the Assist platform

There were several constraints that we needed to be wary of during this project:

  • all the specific, existing use cases that the Assist toolbar needed to fit seamlessly into

  • adhering to existing style guidelines and design language with some wiggle room for all-new AI UI components


Developments along the way included:

  • design all-new marketing graphics and a video to launch the feature

  • pendo guide