
Work / Case 01, ZYOD (ERP)
TNA Mobile Activity Cards
A mobile card that took TNA off the planner's desktop and put it onto the factory floor. 100% adoption in the three factories we launched.




















Planned: 10th Jun, 25 Due: 12th Jun, 25
NSNaimish Srivastava
No comment & activity yet
- Role
- Product designer: component, states, filters, factory rollout
- Context
- The flows and the underlying TNA dashboard were already designed when this was picked up. A senior designer reviewed the work; they weren't in the file with me.
- Scope
- Mobile activity surface on top of an existing TNA dashboard
- Timeline
- Add real timeline (kickoff → factory rollout)
adoption across all three launch factories
factories shipped: 2 Gurgaon, 1 Jaipur
production stages tracked: Fabric Inhouse, Cutting, Stitching, Finishing
The right artifact for a desk. The wrong one for a floor.
ZYOD's TNA module had a Multi Style Table: every style, every stage, delay reasons, comments, history. Right for a planner at a desk. Wrong for the team leads, fabric chasers, and warehouse supervisors the dates actually belonged to. They were on the floor, with no monitor.
Sales Order
before anyone knew
Found when someone next opened the table. By then it had moved through three stages.
A factory-wide view. A lead owning twelve SOs still had to hunt.
The flows and the table were already designed. My job wasn't to reopen the information architecture. It was to decide what a standing worker needs from that data in the next ten seconds.
Three roles, one floor, different questions.
“What's mine, what's red, what I can escalate without leaving the line”
- Owns
- A set of styles across all four stages
- Before
- Verbal updates and periodic desk checks
“Update inspection, attach a photo, ping the next person, in the aisle”
- Owns
- Fabric from inhouse into cutting
- Before
- Knowing status meant physically being there
“A yellow or red signal early enough to replan labor”
- Owns
- Finishing readiness and dispatch
- Before
- Upstream delay arrived as a surprise at the door
Three jobs. Same list. Different questions. That's why I didn't split the product into three apps.
These roles come from the brief and from how TNA was already used, not a fresh round of interviews. Worth confirming against how these teams actually worked.
A table is a reporting object. A card on the floor is a decision object.
“Shrink the table to a phone” fails. Instead I asked every field in the table one question, and the order of the answers became the card.
03/Information hierarchy
Does this change what someone does in the next minute?
Drag or tap a fieldSeven fields from the table. Pick one, then the shelf it belongs on.
the table held everything a planner might need.
The callKeep what changes a floor decision. Put the rest behind a tap.
If a worker has to open every card to know what's on fire, the component has already lost.
delays don't stay inside a department.
The callOne shared list, with role on top as filters.
A private “my tasks” app would make people faster at their own list and blinder to the line. My SOs, Needs Attention, Due this week: accountability without splitting the system of record.
TNA is only as good as the last update.
The callComment, escalate, reassign, and photo live on the card.
If logging a problem takes a separate screen, people won't log it. Report and evidence are the same motion.
Gurgaon and Jaipur don't share a floor plan.
The callDesign against the ERP's data, not a floor plan.
The card was built on the data the ERP already stored per order, so it shipped to all three factories without a rewrite. Customizing per factory would have made every new unit a design project.
A card component that didn't exist in the design system.
Built from scratch, then dropped into the existing ERP. Each piece answers one question a floor worker actually asks.




Status system
- On track / needs attention / delayed, readable before the card is opened
- Color is the triage layer, not decoration

Context on the face
- Style, brand, contacts, and assignment on the card face
- No second screen for the basics









Planned: 10th Jun, 25 Due: 12th Jun, 25
NSNaimish Srivastava
No comment & activity yet
Actions on the card
- Comment, escalate, reassign, and attach a photo on the card
- The audit trail starts at the problem, not in a 6pm recap



Filters on a shared list
- Needs Attention, My SOs, Due this week
- Bulk-select delayed work to escalate, reassign, or comment in one go
Red and yellow wait under “Needs attention.”
Step 1 of 3
The Multi Style Table didn't go away. Planners still need a dense board. The app is the execution layer the table was never going to be.
100% adoption is the number I can stand behind.
adoption across the three launch factories
- GurgaonAdopted
- GurgaonAdopted
- JaipurAdopted
People who had no TNA access unless they walked to a desk now track material, fabric, and SKU status, coordinate across departments, and escalate from the phone, at the moment the decision is happening.
- Status lived on a desktop
- Delays showed up when someone next opened the table
- Escalation was a call or a walk
- Status visible and editable from the floor
- Delays flagged the moment they were seen
- Escalation and photo evidence in one motion
Adoption is the one number with hard data behind it. Floor tools that miss the job get quietly abandoned; this one didn't. I'd rather leave the three above blank than invent them.
What I would not claim.
The card, its states, and the filters. The data model and core flows were locked: I didn't decide which fields existed, only which ones earned a place on a standing worker's first screen.
- 01Floor connectivity: a card that needs a clean network is still fragile in parts of a factory.
- 02Language: English UI on a Gurgaon / Jaipur floor is a real adoption risk as this spreads.
- 03Conflicting status between roles: who wins when the chaser and the lead disagree on a stage.
The problem was never “make TNA mobile.” The company had a source of truth only one job could use. Production speed depends on the other jobs writing into it while they work.
- Which fields change a decision in the next minute?
- Do we optimize for my list or for the line?
- Where does an update have to live so the data stays honest?
- What can stay one component across factories so rollout doesn't become a design project every time?