Overview
The Strata Cloud Manager Access Analyzer is an AI-driven troubleshooting tool that helps IT and network administrators diagnose access and connectivity issues within SASE environments. Through automatic monitoring and an intelligent conversational interface, it lets users quickly analyze why a user, application, or site is experiencing connectivity failures.
The problem
Troubleshooting access issues in a SASE environment is inherently complex. Administrators typically need to collect logs, security policies, authentication records, and network topology details from multiple sources — tedious and error-prone.
Although Access Analyzer was designed to automate access diagnostics, the initial experience introduced friction. A UX audit and quantitative analysis revealed three problems:
- High page visits, low query submissions — many users reached the tool but abandoned it before submitting a query.
- Cumbersome workflow — users had to manually fill out multiple fields in a rigid sequence.
- Steep learning curve — administrators unfamiliar with the query-building structure struggled to generate meaningful insight.
Finding the real problem
After joining this team I performed a UX audit and noted significant usability concerns with the query workflow. The implementation was too rigid — if an admin didn't format the query exactly right, they could not submit it at all.
Quantitative data backed this up. Analytics showed many users navigating to the Access Analyzer page — indicating clear interest — but rarely proceeding to submit a query.
In user interviews we asked users to perform the task in production. One out of six were able to submit a query without guidance. They described it as “too much work for the result.”
The challenge for the design team was clear: how might we streamline query-building so users can quickly ask questions and get answers without frustration?
Target users
- IT administrators — manage user access, troubleshoot authentication issues, and ensure seamless connectivity.
- Network administrators — diagnose network performance and connectivity failures across distributed users, applications, and locations.
The solution
We redesigned the query-building experience around simplicity, guidance, and speed. Users can choose an example question, which loads a guided pre-built query into the main input, or type a natural-language query while the interface dynamically suggests the next expected input — walking them through query creation.
This progressive disclosure approach means users no longer have to parse and reproduce query syntax perfectly in a free-text field. There are no dead-ends, and a complex access query can be submitted in around ten seconds.
Integration with the generative AI copilot
Access Analyzer was then connected to the platform's generative AI copilot, an assistant embedded across the product. Copilot enables conversational querying from any page, so an admin can ask a plain-language question — “Why can't user Alan Larson access Slack from London?” — and receive an answer or guidance. It interprets the query, consults Access Analyzer, and suggests next steps.
This removes the need to navigate to Access Analyzer at all, bringing troubleshooting to wherever the user already is. For users it feels like asking a knowledgeable colleague who instantly retrieves network data. It also lowers the barrier for less-experienced admins, who can describe a problem instead of constructing a query.
My role
- Principal UX Designer responsible for UX strategy and design of this feature.
- Employed a data-driven design approach, working closely with UX Research.
- Facilitated monthly Data Dive-In sessions with the Product lead and UX researcher to review Pendo data, form insights, and generate hypotheses that guided design enhancements.
- Worked closely with Product and Engineering to incorporate actionable workflows.
- Partnered with the UX lead for the AI copilot product to define the generative AI experience and keep integrations aligned.
- Drove awareness, adoption, and consistency across internal teams.
Impact and results
By redesigning the query workflow and integrating generative AI, we turned a struggling feature into a fast, intuitive troubleshooting assistant. The streamlined builder gets admins answers in seconds, and the copilot integration makes troubleshooting accessible from anywhere in the platform.