SIDD: Customer service AI copilot
We help agents excel
with actionable insights
when every second matters
Rated: 4.7 / 5
used across 170+ client programs
project overview
Platform
Desktop, 0-1
My role
Product designer
Duration
2020-2023
Partners
2 fellow designers,
3 product managers,
1 platform success manager,
2 Dev leads, 1 marketing VP, CEO
Key Features
We supports agents with
minimized cognitive load and attention
Voice transcription
Realtime guidance
Task automation
Auto case notes
Sephora 2024 (incl. Q4 holiday ramp)
Business outcomes
Save money
$889K appeasement costs saved
$277K saved training cost
Save time
6.3 minutes per call saved
(36,000 calls)
30% improved email handle time
62% reduced ramp time
5 weeks reduced training time
Empower agents
90%+ adoption
Satisfy customers
16% improved CSAT scores
Increase revenue
2.5x ROI
Design stories
Finding the keys to
frictionless, systematic, and informative UX
story #1
Realtime transcripts and reminder
What does it show
gathers and document voice chat, so that agents can reply better in the first contact (better first customer direct interaction)
How this resonates with my value
I think info-based decision is importat b/c it solves real problem consistently/collectively
agents can see the convo info in writing (ACTION)
agents can track and refer convo context correctly (ACTION)
core info is not lost / understanding is correct despite high volume (IMPRESSION)
How does it help
resolution (RESULT)
Why do people need this (DEFECIENCY, URGENCY)
lost info/misunderstanding is common
human capacity is limited, it’s hard to guarantee consistent results (efficiency with volume)
existing tools are slow and prone to error
first call resolution - volume - contract fulfillment - retention/attraction
Context
Learning the Way Agents Work
It’s more than multitasking. It’s cognitive triage.
Listening
to a customer explain a billing issue
Preparing
a response while navigating CRM fields
Reading
through history and previous case notes
Trying to remember
the five-step process for that one obscure issue
Doing it all
while staying calm, accurate, and on-brand
Biggest challenges
Why User Adoption
Is the #1 Key to AI Success at Contact Centers
What kills adoption
Cognitive load Kills Adoption
noisy info
unfamiliar layout
Interaction Friction Kills Adoption
clicks
tab switching
friction to a routine workflow,
our strategy
the Jakob's Law
Users prefer websites and apps that work like other familiar products
External Consistency
Reduce Cognitive Load
Prioritize Content
Familiarity Breeds Comfort
be systematic
Reduced cognitive load
Mental model and hosting platform
External Consistency
Platform adherence
be frictionless
Reduce interaction friction
Zero Tab-Switching
Search and/or Suggest?
measuring success
What High Adoption Looks Like
96%
auto case note usage
72%
Real-time guidance usage
30s+ per call
immediate handle time reduction
45%+
improved email productivity (EPH)
80%+
overall usage rates across agent groups
story #2
Tasks automation
What does it show
agents can automate particular tasks (order lookup, warranty claim form) one click
How this resonate with my value
Frictionless AI-automated tasks reduces friction from repetitive work, saving human’s time and energy for more judgement-focused job
How does it help
Action: SIDD automates tasks → agents call focus on talking
Why do people need this
Biggest challenges
Why User Engagement
Is the #2 Key to AI Success in the Contact Center
onboarding
feature discovery
troubleshooting
story #3
Autom case note
Why do people need this (DEFECIENCY, URGENCY):
How does it help
agent can finish call logging with minimum efforts
→ agent can work on other ACW like updating database, followup, and submitting product feedback
How this resonates with my value
foundation for long-term support
What does it show (INTENTION UNDER THE PROJECT CONTEXT):
it summarizes convo and highlight actionable items for cross-functional departments to follow up
Business outcomes
save time
Voice-30s+ talk time
Voice-ACW time/%
save money
usage rate; user
hiring cost
refunding cost
make money
email per hour
chat concurrency
Reflection
adoption is not the end of mission
© 2026 Evan Zhai | Designed and built using Framer.