Increased ticket resolution by 20% by building Desk for intakes from different sources

Increased ticket resolution by 20% by building Desk for intakes from different sources

Desk is a standalone, enterprise grade product inside Plane that centralizes incoming requests from multiple sources. It works as a single tray for all intakes like email, forms, in-app, and external tools, allowing teams to triage, route, and resolve requests with clarity and accountability. Desk is not a feature added to projects. It will be a separate, paid product with its own workflow, roles, analytics, and pricing, built for organizations handling high volume, multi team support.

Industry

Project management

Client

Plane

My role

Product designer

Timeline

Apr - Aug'25

Problem & Context

As Plane’s customer base grew, especially with enterprise teams, requests started coming in from multiple channels including email, forms, in-app messages, and external tools like github, discord and even slack. Each channel created work in isolation.

Before Desk, things regularly broke.

  • Constant context switching across tools and projects

  • Missed tickets or requests stuck in the same state for days

  • No clear ownership once a request entered the system

  • Little visibility into where a request was or who was responsible

Project level Intake helped teams manage work inside individual projects, but it did not scale for organizations handling hundreds of requests across customers and teams. Enterprises needed a layer above projects, a place to see demand clearly before execution.

Desk was designed to be that layer.

Goals

  • Provide a single, reliable entry point for all incoming requests

  • Enable fast triage without forcing early commitment to a project

  • Make ownership, status, and next steps visible at all times

  • Scale to hundreds of tickets per day without chaos

  • Stay people‑driven first, automation‑ready later

Users & Roles

Primary users

  • Triage and support teams handling incoming requests daily

Secondary users

  • Desk admins managing routing and escalation

  • Project admins resolving routed work

Buyer persona

  • Founder or Operations Lead

Target audience

  • Enterprise teams with large customer bases and multiple internal projects

My Role

I was the first designer on Desk, partnering briefly with a lead designer before owning the product end to end.

  • Owned UX, interaction, and visual design

  • Translated PM PRD into clear workflows

  • Led rapid iterations with internal teams

  • Designed Desk as a standalone product

The first usable version shipped in one week, with the design evolving over six weeks of iteration.


Research & References

Research inputs came from multiple sources.

  • Internal support and ops workflows

  • Feedback from enterprise customers

  • Existing Plane Intake patterns

  • Competitive analysis of tools like Pylon, Zendesk, Zoho, Jira service management

The key insight was that enterprises do not think in projects when a request arrives. They think in requests, urgency, and accountability. Project assignment comes later.


Core Mental Model

Every request starts the same way.

New → Triage → Resolve or Execute

Desk treats all incoming items as requests first, not tasks. A request earns structure and ownership before it earns a place inside a project.

This mental model guided all interaction decisions.


Key Product Decisions


  1. Board over inbox
    Requests are shown on a state based board to make flow, aging, and bottlenecks visible at a glance. Lists tend to hide issues, while boards surface them clearly. Since most teams are already familiar with Kanban, a left to right board makes it easier to understand how a ticket progresses compared to an inbox or list view.

  2. Single entry state
    All requests enter Desk in the same New state, regardless of source. This removes ambiguity and gives triagers a predictable starting point.

  3. Explicit project assignment
    Routing to a project is always a deliberate action. This prevents silent misrouting and makes accountability clear.

  4. Global identity preserved
    Desk items keep their identity even after being assigned to a project, ensuring traceability without forcing teams to change how they work.

  5. Manual first by design
    Phase 1 avoided automation intentionally. With high volume, correctness mattered more than speed. Automation was designed as a follow up layer.

  6. State driven interactions
    Actions are tied to states instead of generic controls. Details are progressively revealed, and triage is clearly separated from execution. This keeps cognitive load low even at high ticket volume.


Key Features

Centralized Desk Board

  • Single view of all incoming requests across sources

  • Clear states showing current ownership and progress

  • Filters for priority, assignee, customer, and aging

Triage & Ownership

  • Desk admins can comment, resolve, or route requests

  • Ownership is always visible

  • Requests never disappear into projects without traceability

Project Routing

  • Requests are assigned to the right project when ready

  • Project teams keep their existing accept or decline workflows

  • Status updates reflect back into Desk

Source Context

  • Each request retains its original source

  • Content and attachments remain visible during triage


Constraints & Trade‑offs

  • No automation in Phase 1 despite high volume expectations

  • Manual routing to ensure correctness and trust

  • Separate permission models between Desk and Projects

  • Limited cross‑tool sync to avoid partial or misleading states

These constraints helped keep the system understandable while laying the foundation for future scale.


Outcome & Early Signals

Desk is currently under development and not yet fully shipped, but early internal usage and simulations indicate strong potential impact.

  • Anticipated reduction of 25–35% in missed or stale tickets

  • Estimated 2–3 hours saved per support agent per week through reduced context switching

  • Faster first response due to a single, visible queue

  • Clear accountability leading to fewer tickets stuck in one state for long periods

These numbers are directional and based on internal workflows, early testing, and enterprise usage patterns.


Learnings

  • Centralization only works when ownership is explicit, and when visibility is prioritised over predictability

  • Products become usable through continuous iteration with users and product owners, while keeping business goals in focus

  • Strong products require taking opinionated decisions and being able to justify them clearly to stakeholders

  • Even during development, data played a key role in guiding decisions, especially when designing for scale and a wide range of users

  • Designing on top of existing tools requires restraint and respect for established workflows

Desk reinforced the importance of designing for flow and accountability, not just capture.


What I’d Explore Next

  • Rule‑based automation for routing and prioritization

  • Role‑based and group‑based flows as RBAC and GAC mature

  • SLA views and aging analytics

  • Customer‑facing, member‑only portals to track request status

These would help Desk scale from a strong operational foundation into a system that actively manages demand.