STLabs Standard Template Labs

The deepest view of a ticket — and where every decision starts.

Designing the individual ticket detail: a split-view experience where AI agent conversations meet structured ticket data, across any source system.

RoleUX Designer
Timeline2025
Type0→1 Product
FocusMulti-source Data, AI Collaboration

The problem

Enterprise ITSM tickets come from everywhere — Slack, ServiceNow, Jira — each with different data and context. Admins context-switch between systems to understand what's happening and figure out what to do. No single view synthesizes what the user reported, what the system knows, and what action to take.

Enrichment

The hero content.

By grounding every ticket our graph database, I replaced the traditional description with a synthesized brief that merges user-reported issues with live environment context. This layout uses a "We found" transition to clearly distinguish human input from AI reasoning, guiding the eye toward bolded critical facts that drive faster triage.

Key insight

Conversation sources (Slack): enrichment absorbs everything — no separate description. Ticketing sources (ServiceNow, Jira): the original form submission appears separately as reference. The enrichment still synthesizes, but the source material is preserved.

Interactive Click to compare

Slack ticket (enrichment absorbs everything) vs ServiceNow ticket (description preserved separately).

Workflow lifecycle

The agent doesn't just suggest — it executes.

Confidence through clarity. The workflow run isn't just a status—it's the core utility of the product. I designed the lifecycle to be calm and deterministic, moving from suggestion to execution without unnecessary friction.

Design principles

Brand as signal. The STLabs logo pulses in place of the Run button. No spinners, no "Running" text—just the brand doing the work.

Deterministic failure. Transient errors offer a Retry. If that fails, or if the error is terminal, the system auto-suggests an alternative or opens a chat to investigate.

Closed loop. Completion cross-fades the action block into a persistent resolution record, signaling that the work is truly done.

Interactive — Happy Path
Okta MFA Reset

Unlock account and push MFA re-enrollment to Grant Fuhr.

Interactive — Retry & Auto-Suggest
Restore Workday Access

Re-enroll MFA device and restore Workday access for Sarah K.

Activity timeline

The record of everything.

Ticket closed 11:14 AM PST
JK
Jordan K. ran Okta MFA Reset 11:02 AM
Agent assessed ticket 8:42 PM PST
Suggested Okta MFA Reset · Assigned to Jordan K.
Show less
Queried Okta auth logs — account locked, 3 failed sign-ins from 10.0.47.12
Queried Okta user profile — MFA token expired, last enrolled device replaced
Ticket created via Slack 8:41 PM
↑ newest    oldest ↓

The timeline has three actors: Source (where the ticket came from), Agent (our agent), and Human (any person). Plus outcomes — results with no actor.

The spine icon always shows who. The text shows what and where. Newest events first. Collapsed by default with a chevron and event count — the enrichment and suggested resolution above are the primary content.

The full picture

Table to conversation in one click.

The split view brings it all together — agent conversation on the left, ticket detail on the right. The admin reads the enrichment, clicks Run, watches the timeline build, handles complications through chat.

Ticket table → split view. Agent conversation on the left, ticket detail on the right.

Explorations

Roads not taken.

Interactive Click Here

Per-field blue labels (noisy) and Horizontal metadata strip (competed for vertical space). We moved away from these early approaches because they either created too much visual noise or restricted space for critical enrichment content.

Other iterations: "Context" → "What we know" → "Summary" → no label. "Checked" → "Sources." "Agent processed ticket" → "Agent assessed ticket." Each round shaped by whether the language described the system's action or the user's mental model.