Planning & Scheduling
Innovapptive · Enterprise product
Making complex maintenance work easier to plan.
Four connected workflows that help maintenance teams prepare jobs, assign people, and work around equipment downtime.
- MY ROLE
- UX/UI, interaction design & frontend prototyping
- PERIOD
- April–September 2026
- TEAM
- Product manager, VP of product, design director & two engineers
- STATUS
- Working prototype for stakeholder review
A working frontend connected Planning, Scheduling, Downtime Calendar and Plant Maintenance Optimization. Stakeholders could review decisions in context, beyond static screens.
The challenge
Prepare the work. Then find the right time and people.
A maintenance job needs more than a date on a calendar. Someone prepares the scope, parts, instructions and prerequisites; someone else decides who can do it and when. Those decisions need to stay connected without becoming one overwhelming workspace.
I mapped the responsibilities through requirements, product discussions and stakeholder reviews. The team consulted in-house industry experts and existing clients. My focus was the handoff between preparing work and building a workable schedule.
Planner
Prepares the work package: scope, operations, parts, hours, permits and isolation requirements.
Scheduler
Assigns released work to technicians and time slots, considering capacity, readiness and equipment downtime.
01 · Prepare
Make the next step in preparation explicit.
Draft work stays in Planning. Where required, Field Validation adds an on-site check before release. Ready for Schedule moves the work into the scheduling backlog. Jobs that do not need a new work package can go directly to that backlog.
I used shared table patterns across Planning and Scheduling so the same work remains recognizable as it moves between roles. Manual preparation and AI-assisted planning feed into the same reviewable work package.
02 · Schedule
Design for screen height, not just width.
A June review exposed a practical problem: with four panels open, smaller laptop windows showed only one or two rows of work. More information on screen had made the actual schedule harder to use.
I reduced toolbar space and used compact, 32-pixel rows with ID-first labels. The team continued with this revised workspace. A backlog card rail remained an explored alternative rather than the final direction.
Keep readiness visible
Missing materials, permits and isolation remain visible without preventing a future slot from being reserved.
Handle date exceptions separately
Scheduling outside the basic date range asks for an explicit choice: cancel or schedule anyway.
03 · Review AI
Put recommendations beside their consequences.
The earlier Auto Schedule concept exposed agent controls. Stakeholder feedback shifted the focus to the scheduler’s decision: what is proposed, why, and what changes on the board?
I moved recommendations beside the schedule. Schedulers can inspect related work orders, review bundling and retract choices before accepting a proposal. Accepting a proposal is separate from freezing a week for ERP transfer.
04 · Connect
Give different decisions their own place.
For the Downtime Calendar, I moved summaries from above the calendar into a side panel. Dates, events and downtime hours can now be read together; summaries follow the visible dates and filters.
Plant Maintenance Optimization became a separate review queue. Reviewing duration and frequency recommendations has different owners and timing from preparing a work package. Planners compare current and proposed values with supporting records before sending a recommendation to Master Data Management or rejecting it.
Outcome & reflection
A prototype people could actually review.
I directed AI-assisted implementation in Angular using CWP components and Claude Code, reviewed the interactions and verified the visible result. The frontend made it possible to review real flows with consistent demo data.
ERP and Master Data Management handoffs are simulated. This is stakeholder-reviewed prototype work, with no measured business impact claimed. Task-based sessions with planners and schedulers are the next validation step.
My main learning: follow the ownership of a decision, reuse interactions across related workflows, and check the available screen height early. A screen is only useful if people can see enough of their work to make a decision.
