role
Product Designer
type
Enterprise, GenAI, Responsive Toolbar
MY CONTRIBUTION
AI Product Design, AI- Assisted Prototyping, Interaction Design, Cross-functional Collaboration
Foundever
Designing AI-assisted wrap-up for customer support agents
Designed a practical in-call copilot that delivers real-time next-best-action recommendations and compliance guidance to customer support agents, driving 2× growth in tool adoption
5 months
·
2026
·
Team: 2 Product Designers (me+1), PM, QA, AI Engineers, UX Researchers, Sales
,
3.98M+
Agent hours saved annually
2x (55,000 → 120,000)
,
Increase in tool adoption rate
-52% (2.5 min → 1.2 min)
Avg. wrap-up time
DISCLAIMER
🔒 This project is under NDA
I am happy to discuss my design process, technical decisions and collaboration during interview!
OVERVIEW
55K agents were provisioned with the AI copilot — but only 18% agents used it.
The tool is Foundever's GenAI copilot for CX agents, real-time guidance, automated summaries, instant answers. Only 18% used it. Because it would not solve the actual user problem of helping them
I redesigned the experience layer — making the AI's intelligence usable and trusted by front-line agents in high-pressure, real-time environments.
The business goals were clear: Improve product adoption and ultimately improve agent efficiency
PROCESS
Shadowed the agents in their working environments
03
Field visit
10
User Interviews
1000
User surveys
Agents needed a low-friction in call way to guide them in real time
The tool automated some parts of the process like write, translate but the wrap-up feature wasn't adding much value :
Compliance steps were easy to miss
Customers did not understand why some trees would ship months after purchase
Personal templates were common workarounds
Hardiness zone information was buried and was not surfaced as a confident purchase signal
Inconsistency in solving cases
Basic filters helped narrow products, but did not explain which tree was right for a specific yard or climate.
Technical Constraints
zero data retention
PII redaction
VDI fallback
How might we design agent experience with minimal disruption to agent flow?
Design solution
Redesigned the experience to make the AI approachable and trustworthy through human-in-loop experience by surfacing real-time compliance and next-best-action recommendations.
01
Surface real-time compliance
02
Next best action during call
03
Build confidence into the product experience
Trade offs
latency vs freshness
automation vs agent control
Wrap-Up & Compliance
Real-time guidance, every call guides agents through compliance steps, flags language violations, and generates a call summary ready to paste into CRM notes. The tool surfaces the next best action, agents stay ahead without losing momentum.
IMPACT
Team member feedback
"It helps me get through call wrap-up faster while keeping me in control of the final response and is easy to use"
Support agent, Week 8 post-launch