Enterprise task infrastructure for Exchange Platform—centralizing triage, ownership, and auditability inside the ops system teams already run.
I led discovery with ops and management, mapped the task lifecycle end to end, and scoped Phase 1 around scalable triage, clear accountability, and enforceable closure—not feature breadth.
ResultImproved task visibility, operational consistency, and workflow scalability across distributed teams.
UX Research · Product Design · Workflow Design · Internal Tools
Teams executed in Exchange while coordination happened in Slack, email, and spreadsheets—unclear ownership, manual tracking, and inconsistent prioritization across high-volume triage.
Research mapped how tasks entered the business, how teams coordinated under load, and where Exchange could enforce workflow rules without breaking throughput—grounding every Phase 1 decision in observed ops behavior.
I mapped the full Task Center lifecycle—from creation through reporting—to pinpoint where coordination broke down. Exchange made transactions visible; the work to resolve them lived elsewhere.
Capture work from orders, communities, or customer contact.
Logged late in Slack or sheets—often unlinked to Exchange.
No standard intake; formats varied by team.
Duplicate entries; fear of losing requests.
In-context creation tied to operational records.
Route work to the right owner or team.
Unclear ownership; tasks sat unassigned.
Manual routing through chat—no shared queue.
Stalled work triggered finger-pointing.
Visible assignee and authoritative ownership.
Order daily work by urgency, due date, and volume.
Priority language differed by team and shift.
Managers exported lists to rank workload.
High load deciding what to tackle first.
Tabbed queues, due dates, scannable rows.
Execute triage against operational records.
Parallel trackers because Exchange had no queue.
Full-page hops broke sequential triage.
Duplicate follow-ups at peak volume.
Inline detail panel—queue stays in view.
Communicate progress to managers and adjacent teams.
Status lived in chat; labels meant different things.
No single status model across teams.
Standups surfaced distrust in reported progress.
Enforced To-Do → Assigned → In Progress → Completed.
Verify resolution before closing or escalating.
QA in side channels; criteria inconsistent.
Delayed approvals when managers were unavailable.
Rework when closure standards misaligned.
Structured review with comments and shared rules.
Mark work done and release from the active queue.
Silent edits in informal trackers—no history.
Disputed ownership with no audit trail.
Work reopened without record of who closed it.
Locked completion, controlled reopen, activity log.
Report throughput, backlog, and stall patterns.
Manual rollups from sheets and Slack.
No visibility without exports and outreach.
Leadership trailed floor conditions by days.
In-platform queue metrics; Phase 2 reporting base.
Coordination failed not because teams lacked discipline—it failed because Exchange showed transactions without showing the work required to resolve them.
Synthesized from interviews, contextual inquiry, and workflow mapping—these patterns directly informed Phase 1 scope and system rules.
“If I don't write it down the second I hear it, it's gone. Nothing in Exchange tells me what I still owe by end of day.”Operations specialist, contextual inquiry
Each insight maps to a shippable workflow rule—structured enough for accountability, flexible enough for real triage under load.
Operators maintained parallel trackers because Exchange had no system of record for open work.
Task Center module beside Dashboard, Orders, and Payments—tasks tied to operational records.
Product outcomeSingle queue for creation, assignment, and updates—less cross-tool reconciliation.
Managers could not scan stalled or unassigned items without exports or direct outreach.
Status tabs plus assignee and status columns in the list row.
Product outcomeFaster coordination across teams during standups and peak periods.
Full-page navigation per update broke sequential triage throughput.
Inline detail panel for status, assignee, and comments—queue stays in view.
Product outcomeReduced context switching in workflow shadowing sessions.
Disputed completions had no durable history in informal trackers.
Locked completion, explicit reopen warning, permanent activity log.
Product outcomeClear escalation path when ownership or timing was challenged.
Phase 1 modeled one operational path—from intake through resolution—so every screen reinforced the same system rules before UI detail.
Work enters from Exchange context—linked to orders, communities, or payments.
Owner assigned; task surfaces in shared queue with assignee visible.
Tabs, due dates, and list scan support daily triage order.
To-Do → Assigned → In Progress → Completed with consistent labels.
Comments, locked completion, controlled reopen—activity log preserved.
Three modules mirror how ops triages at scale—list scan, in-flow updates, enforceable closure—using Exchange patterns operators already know.
01 · Queue at scale
02 · In-flow updates
03 · Trusted closure
Measured against pre-launch fragmentation—not vanity feature counts. Signals the ops team could observe in Phase 1.
+100%
Task Center usage vs. prior scattered trackers (post-launch, internal reporting).
One status model and assignee field across teams—onboarding aligned to tab semantics.
Fewer Slack/sheet loops to confirm “who owns this” during peak triage.
Same queue architecture ready for bulk actions and SLA signals in Phase 2.
Enterprise internal tools fail when they optimize for configurability over adoption. Here, structure had to be firm enough for accountability—locked closure, visible assignee, status tabs—while leaving room for how teams actually triage under load. Edge cases (reopen, disputed ownership, empty queues) shaped as much of the design as the happy path. The systems bet was simple: make Exchange the coordination layer, then scale rules in Phase 2 without asking operators to learn a second product.