AI Support in CX Cloud
Role
Product Designer
Company
Cisco
Tools
Figma
Project Overview
Cisco’s CX Cloud, a cloud-based network security and customer experience platform, faced case overload due to both critical and minor issues, causing delays and straining resources. Users and the Technical Assistance Center (TAC) engineers sought better proactive remediation and self-help options to resolve problems faster and reduce overall support case load. This project focused on designing AI-driven tools to empower users and ease TAC’s workload.
Goals
Empower users with early detection and actionable AI-guided remediation to resolve common issues independently.
Reduce avoidable support cases and alleviate TAC workload, allowing engineers to focus on complex problems.
Build user trust by creating a scalable, accessible support experience that integrates smoothly with CX Cloud.
Research
To understand the landscape of existing solutions, I conducted a competitive analysis of self-help and remediation tools across leading enterprise platforms. The goal was to identify best practices, usability patterns, and gaps that could inform a scalable, AI-driven support experience. In parallel, I gathered insights from TAC engineers and internal support stakeholders.
Key insights:
Top-performing tools surfaced support content contextually, reducing the burden on users to search manually
Some platforms used AI to recommend remediation actions, but often lacked transparency in how suggestions were generated
AI guidance should be not only actionable, but also understandable and trustworthy to encourage adoption
There is a clear opportunity to build user confidence by linking AI-driven recommendations to existing Cisco documentation.
Design
I led the end-to-end design of these experiences in Figma, from early concepts to high-fidelity prototypes.
I began with rapid sketches to explore key flows and quickly align on direction.
Once the best ideas emerged, I translated them into detailed wireframes to refine the layout and interaction before moving into high-fidelity design.
Final Designs
Below are key screens from the final prototypes showcasing the Predictive Remediation and AI Self-Help Center features.
Predictive Remediation
Surfaces predicted incidents and remediation plan options to avoid a potential crash
Offers guided remediation steps with plain-language explanations
Allows users to act directly or explore linked support materials
AI Self-Help Center
Asks users to describe their issue in their own words
Uses AI to analyze the issue and recommend relevant solutions based on input quality (the more detailed the description, the stronger the recommendation)
If suggestions don’t resolve the issue, users can open a case directly from the same interface
Case creation is streamlined with fields pre-filled based on the original issue description
Testing & Validation
At the time of this case study, the product had not yet launched and no formal user testing was conducted. However, internal stakeholder feedback from TAC, product, and design teams indicated strong support for the proposed solution and its potential benefits.
Impact
Business analysts estimated that improving self-help and proactive remediation could reduce Cisco’s overall support case volume by 3%, saving approximately $15 million per year in support operations costs.
“Improving self-help could save Cisco up to $15M per year.”
— Business Analyst, Cisco Customer Support
The approach demonstrated strong potential to reduce costs, scale support, and improve the user experience, even before full deployment.
Lessons Learned
Transparency in AI suggestions will be critical for user trust and adoption
Proactive support should reduce friction, not add steps
Collaboration across TAC, engineering, and design leads to more grounded, scalable solutions
Next Steps
Conduct usability testing and continue refining based on feedback from early users
Collaborate with engineering to implement AI detection and remediation logic
Expand to support additional use cases and user roles
