Cisco HelpZone

Redefining Support with AI

Redefining Support with AI

HelpZone is a internal employee support center. Home to a knowledge base spanning IT and People & Culture (HR) resources, with options to open cases or reach live support. Despite its breadth, more than 75% of employees bypassed it entirely, turning to colleagues or direct support channels instead. I led the full experience redesign across Cisco’s homepage, help topics page, and search results page, integrating AI throughout the flows to improve the employee support experience.

Role

Lead Product Designer

Platform

Web

Year

Spring 2025 - Fall 2025

Status

Shipped

Disclaimer: This is a brief overview of the work done. For a thorough case study please contact me :)

KPI

Business Impact

0%

0%

increase in self
resolution confidence

0%

0%

lift in tile adoption

0%

0%

improvement in case
tracking efficiency

0%

0%

increase in attempting
before contacting an agent,

Context & Problem

What is HelpZone?

HelpZone, Cisco’s internal IT and HR support platform, faced a major trust and self-service problem.

Despite its broad resources, over 75% of employees completely bypassed it, opting to ask colleagues or go straight to direct support.

How the problem came to light.

"I can't find
the support channels!"

HelpZone received feedbacks that users were having difficulty finding live support or case submission pathways.

In a 20-user study with UX Research, 16 participants failed to scroll far enough to see the channels—meaning they weren't missing, just out of sight.

16/20 Users reported not scrolling all way to the bottom to find the support channels

16/20 Users reported not scrolling all way to the bottom to find the support channels

16/20 Users reported not scrolling all way to the bottom to find the support channels

What the users wanted to see.

What the users wanted to see.

What the users wanted to see.

But stakeholders thought differently.

The Hurdle

Stakeholders told us that they were cutting live agent support by 2/3 next quarter

So we asked the question….

Why are users looking for live agents in the first place?

So we reframed it. New Core Problem…

How might we help employees find solutions to their problems through HelpZone, quickly and on their own?

Focus Areas

I pulled our behaviorial data to understand where users were currently clicking for self help. I priortized top areas of clicks and traffic.

Where everyone starts

Homepage

Every user's journey of self-help and resolution begins on the homepage. What shows on the homepage determines what path user will take.

29% of clicks

Search Results Page

Where the user initiates their self help by clicking on the search bar and landing on the search results page.

12% of clicks

Help Topics Flow

Users click a help topic tile to access a curated article page but it was underperforming. Data suggest users are confused by the long, uncurated lists

Homepage

Data driven design decisions

HELP TOPICS

I used PowerBI search data to dynamically surface the most popular categories on the homepage.
By placing the most-used tiles first, employees can instantly find what they need without scrolling or guessing.

Device Experience
Network Access
Buying & traveling
Home & Remote Access
Pay & Compensation
Time off
Exits
People Data & Tools
Device Experience
Network Access
Buying & traveling
Home & Remote Access
Pay & Compensation
Time off
Exits
People Data & Tools
0K
2K
4K

Popular articles clicked via Search/Typeahead

Device Experience
Network Access
Buying & traveling
Home & Remote Access
Pay & Compensation
Time off
Exits
People Data & Tools
Device Experience
Network Access
Buying & traveling
Home & Remote Access
Pay & Compensation
Time off
Exits
People Data & Tools
0K
2K
4K

Popular articles clicked via Search/Typeahead

Device Experience
Network Access
Buying & traveling
Home & Remote Access
Pay & Compensation
Time off
Exits
People Data & Tools
Device Experience
Network Access
Buying & traveling
Home & Remote Access
Pay & Compensation
Time off
Exits
People Data & Tools
0K
2K
4K

Popular articles clicked via Search/Typeahead

Used those categories to structure the homepage as "Help Topic" tiles

FEATURED ARTICLES → SPOTLIGHT MODAL

Less than 1% of users clicked the homepage's "Featured Articles" section.

Yet IT stakeholders still required it, taking up real estate that could serve a better purpose.

BEFORE

Less than 1% of users clicked the homepage's "Featured Articles" section.

Yet IT stakeholders still required it, taking up real estate that could serve a better purpose.

BEFORE

Less than 1% of users clicked the homepage's "Featured Articles" section.

Yet IT stakeholders still required it, taking up real estate that could serve a better purpose.

BEFORE

AFTER

AFTER

AFTER

To maintain consistency and reduce development time, I proposed reusing the Spotlight modal from our Bridge employee platform for "Featured articles." I worked with engineering to personalize these articles by user location and added a discreet "contact support" link to an article with phone numbers. While this didn't display the support channels exactly as the user requested, it successfully preserved access to live support.

Help Topic Tiles Flow

Making Help Topic tiles useful.

The goal of this work was to improve the employee support experience across the homepage, help topics, and search.

Sub Topic
Region
Country
Sub Topic
Region
Country
000
200
400
600
800

Filter Usage

Based on sidebar engagement, we removed the filter panel and used subtopics to restructure the Help details page.

Sub Topic
Region
Country
Sub Topic
Region
Country
000
200
400
600
800

Filter Usage

Based on sidebar engagement, we removed the filter panel and used subtopics to restructure the Help details page.

Sub Topic
Region
Country
Sub Topic
Region
Country
000
200
400
600
800

Filter Usage

Based on sidebar engagement, we removed the filter panel and used subtopics to restructure the Help details page.

A/B Testing

A/B Testing

A/B Testing

I partnered with a UX researcher to test two navigation directions with 14 users. Direction A had a 43% misclick rate and an average task time of 19.6 seconds, while Direction B had zero misclicks and averaged just 4.4 seconds. This validated that users favored simpler navigation, even with an extra step.

After

After

After

Reflection

After thoughts

What I learned

What I learned

What I learned

←

Explore more work