Indexer Status
Replacing a failing status page with chat-guided troubleshooting that decreased support tickets.
Role
Product Designer
Team
1 Product Designer, 30 Engineers, 2 Product Managers
Tools
Figma
“Since this product is so complex, the idea of using a chatbot to assist users is revolutionary. I have not seen any other product do this. And it should be in every one of our products.”
The Problem
A status page that nobody used
The product team repeatedly received customer feedback that the indexer status page was not meeting their needs. Users were frustrated that they weren't being notified of issues, and it was difficult to see problems and then manually investigate them.
Problem Statement: System administrators are failing to keep the indexing system operational due to a lack of information on the status of the indexing service, causing frustration and risk from sessions not being indexed.
Research
Understanding the problem through structured research
Persona: Steven System
Direct customer access was limited, so I created Steven System — a persona representing the system administrators who rely on this page. Steven helped the team empathise with users' goals and share a consistent understanding of who we were designing for across engineering, product, and support.
Jobs to Be Done
I ran workshops with PMs, sales, and support to capture customer tasks in a structured way. This revealed hidden requirements and three distinct categories of tasks: urgent errors needing immediate attention, system health at a glance for future planning, and performance information for long-term maintenance.
Understanding the existing design
I spoke with subject matter experts and support to understand how users currently used the page. The feedback was clear: many users didn't use it at all because it provided no real value. It didn't surface useful information, didn't help them maintain the indexer, and didn't provide errors, solutions, or prevention.
Design Principles
Four principles to guide the redesign
- Show priority quickly — colourise items (while considering accessibility).
- Educate consequences — help users understand the impact of issues.
- Guide to a solution— don't just show problems.
- Design modular — accommodate future iterations.
Key Decisions
From showing problems to solving them
Three categories of tasks
Research revealed three distinct task categories: urgent errors/warnings needing immediate attention, system health at a glance for future planning, and performance information for long-term maintenance. Due to technical limitations, only urgent and status sections were included in the first release.
Design Challenge #1
How do we manage cases with a large number of alerts — without scroll-in-scroll, empty space, or content shifting?
By placing errors and warnings at the bottom of the health status section, the design achieved one scroll for all content, no shifting, and no limit to the number of items. I was careful not to create technical debt by restricting future content additions at the top of the page.
Design Challenge #2
How can we make it clear and intuitive that some errors are related to a specific health status?
Hovering a health status tile increases its elevation. Clicking filters the errors/warnings to show only related items — without shifting content. I also added hover states on related alerts when hovering a health tile.
A new problem surfaces: showing problems isn't enough
During ideation, it became clear that simply presenting problems didn't improve the experience. Users had to manually find the problem, search a knowledge base for a matching issue, then follow instructions. This was costing the business money in support tickets.
Additional problem statement: System administrators struggle to find solutions to indexing issues, resulting in frustration and an increased risk of system instability.
This led to the most innovative part of the design — a chat-guided troubleshooting experience that highlights errors and guides users step by step to a solution, directly within the product.
Final Design
Chat-guided troubleshooting
Testing
Adapting to how users actually work
During internal testing and prototyping, interesting interaction issues arose. Users often preferred having many windows open simultaneously — we couldn't anticipate the size or number. I added:
- The ability to move Chat into an external window.
- Automatic return of the external Chat window if closed.
- Auto-minimise Chat when the window gets too small, with a prompt to move to an external window.
- Links clicked in the external Chat window affect the source window.
The Outcome
A design that changed how the company thinks about support
The redesign shipped and was tested with real customers to very positive feedback. But the biggest impact went beyond this single page.
The chat-guided troubleshooting approach triggered a company-wide conversation about how One Identity delivers support. The traditional model — users search documentation to find solutions — was replaced by a new vision: the product anticipates problems and guides users to solutions conversationally. This design didn't just fix one status page — it established the direction for how the company approaches support across its entire product suite.
As Richard Hosgood put it: “It should be in every one of our products.”
