AI Agent Assistant

One week product innovation workshop

GENESYS

Despite being defunded, the Product Manager wanted to show how Genesys Agent Assist could evolve into something uniquely valuable to the company. I collaborated with her engage the UX Research team and facilitate a ideation workshop with product and AI teams to understand agents’ current experiences and visualize a future state.

Impact

Immediate funding approval | prioritized by leadership post-workshop, despite initial de-funding.

Evolved into Genesys Agent Copilot | started the foundation of agent workspaces, knowledge management, and next-best action capabilities.

2022 Ventana Research named Genesys the 3rd top (out of 18 companies) in agent management optimization

“This was too good to not fight for and Winter did a great job leading us there.”

— Agent Assist Product Manager

1-Week Product Vision Workshop

The workshop took place over 6 sessions (2 weeks) following the completion of discovery UX research study and participants were invited based on their expertise and skill sets. We carefully defined the problem statement and agenda. Additionally we gathered, what I like to call, "inspiration fuel” which included user research, market research, and comparative analyses of innovative companies.

Problem Statement

By the end of 2022, how might we leverage AI to create “super agents” who can

provide more personalized customer service
create additional value for the business
apply ever-green practices and knowledge to their service

while also reducing their cognitive load?

Workshop activities

Participants

Research & SME readout
Needs statements
Ideation (x2)
Prioritization
Ideate capabilities
Experience-based roadmap
Storyboards

3 | Product Managers

1 | Data Scientist

2 | AI Engineers

1 | UX Researcher

2 | UX Designers

Research insights

The UX Research helped us identify that we needed to provide agents

a way to learn best practices more quickly remotely
correct and up-to-date knowledge at their finger tips
easier ways to understand and get to desired customer intent and outcomes
more direct ways to support business goals

Solution ideation

After drafting needs statements, the participants ideated over two sessions. They were tasked with find inspiration following the first session to iterate on drafted ideas. Originally the “Pinterest” idea was prioritized, but after further discussion the AI Engineers supported the “Waze” concept and believed it was feasible within our desired timeline.

Concept 1 | “Waze” AI Assistant

A bot guide agents toward delivering a customer outcome by combining contextual data. It will tell an agent it’s current hypothesis, how they’ll get there and provide multiple path options if available.

Concept 2 | “Pinterest” for Agents

An agent can curate a digital collection of useful information from any content (voice, text, images, screenshots, etc) without taking focus away from their customer conversation nor slowing down.

Roadmap and Storyboards

Participants ideated further on potential capabilities and we broke them down by primary jobs-to-be-done and hypothetical phases of delivery. They then iterated to fill in gaps and storyboarded out the concepts. The engineers recommended that we prioritize the “Pinterest” concept for a first release to give them time to build out a technical proof of concept for the “Waze” concept.