Project Specs

Delhivery HRMS (HR Management System)

Delhivery is one of India’s largest logistics and supply chain companies, powered by thousands of people working across warehouses, delivery hubs, and sorting centers. Many of these people are sourced through third-party vendors.

But the systems supporting this workforce were anything but streamlined.

Across regions and vendor types, processes like requisition sharing, onboarding, payouts, and employee tracking were being handled through a patchwork of tools, emails, and WhatsApp groups with little visibility and zero consistency.

My role was to help design a centralized HRMS ecosystem.

We created two purpose-built portals:

• A Manpower (Vendor Portal) to empower manpower vendors with data, visibility, and ownership

• An Workforce (Employee Portal) to unify internal HR and Ops workflows, from requisition to payout

After rolling out key modules, vendors reported reduced dependency on regional HR teams, faster access to payout data, and improved retention during onboarding. Internal teams reported higher accuracy in capacity planning and decreased grievance resolution time.

My Role

Research
Conceptualisation
Design

Usability testing

Dev handoff
Design Sanity

Team

1 Product Designer
1 Product Lead
3 Product Managers
2 Engineering Leads
18 Engineers

Duration

Phase 1: Nov'24 - Feb'25

Phase 2: Mar'25 - Present

Context

Imagine a pile of oranges

Imagine a pile of loose oranges rolling around on a table. They are all important, but without a basket to hold them together, they bump into each other, scatter, and sometimes go missing.

That was the state of vendor and HR operations at Delhivery.

As one of India’s largest logistics companies, Delhivery relies on thousands of workers across warehouses and delivery hubs. Many of them are supplied by third-party manpower vendors, who handle different categories such as:

• Offroll workers for long-term roles

• LMA (Last Mile Agents) for commissioned deliveries

• Adhoc or Loaders for short-term, high-volume needs

Despite the scale, critical processes like requisitions, onboarding, and payouts were handled through a mix of emails, spreadsheets, and WhatsApp groups. Each vendor type used a different system.

Our goal was to bring these operations into one basket.

BharatPe's diverse applications faced design inconsistencies and usability issues, leading to a fragmented user experience and reduced engagement. Creating a unified design system was crucial.

Core Problems

❌ Disconnected systems and inconsistent workflows

❌ Lack of real-time visibility across the employee lifecycle

❌ Fragmented and opaque payout processes

❌ Offline, manual communication causing delays and errors

❌ No performance or planning intelligence

BharatPe's diverse applications faced design inconsistencies and usability issues, leading to a fragmented user experience and reduced engagement. Creating a unified design system was crucial.

Proposed Solution

Two connected platforms: Manpower for vendors and Workforce for HR and Ops.

Together, they'll streamline requisitions, onboarding, and payouts across Delhivery’s distributed workforce.

BharatPe's diverse applications faced design inconsistencies and usability issues, leading to a fragmented user experience and reduced engagement. Creating a unified design system was crucial.

Phase 1

To manage timelines and delivery pressure, I divided the project into two phases.

Phase 1 was about speed as I had to get working flows in place quickly. The product team needed working flows in place quickly so the focus was on unblocking development and building the foundation. There was no time for formal research or detailed scoping so I jumped straight into design using past experience, pattern libraries, and intuition.

Had to catch what was rolling ->

I started designing while still understanding system complexity and user behavior. There were constant context switching between teams and features. I juggled multiple live modules with overlapping timelines, to push first screens quickly to unblock development.

✅ What came out of this phase

A working first version of key modules
A foundational design system to build on later
A deeper understanding of internal workflows and tech constraints
A clear sense of what wasn’t working (which helped inform the next phase)

BharatPe's diverse applications faced design inconsistencies and usability issues, leading to a fragmented user experience and reduced engagement. Creating a unified design system was crucial.

Stopped and started sorting ->

Once the initial rush settled, it became clear that the system wasn’t just missing features but clarity. Feedback from leadership called out UX gaps. PMs stepped back and I leaned in.

• We went back to the drawing board

• I re-mapped workflows from scratch and reconsidered each screen

• This time, I had real understanding of user needs and system limitations

• We simplified complex flows and trimmed the unnecessary

• Design decisions were now grounded in purpose, not pressure

BharatPe's diverse applications faced design inconsistencies and usability issues, leading to a fragmented user experience and reduced engagement. Creating a unified design system was crucial.

Phase 2

With the foundation in place, Phase 2 focused on refinement.

I asked five key questions to uncover what truly mattered and designed around those answers, simplifying every screen with intent.

Q1: How to balance info with UX?

When we first tested our tables, users said they struggled to identify which numbers mattered most.

PMs asked for more information even if that meant lots of columns and numbers. I advocated for less clutter and better user experience.

We ran multiple prototypes with different layouts, ranging from flat tables to expandable drawers. By testing comprehension time and error rate in key workflows, we landed on a format that supported clarity and depth.

Q2: How do we help users quickly understand what needs attention?

Our team realized that while data was present, users didn’t know what to do with it. There was no clear call to action or status cue.

To fix this, I:

Mapped user journeys to identify friction
Defined each screen’s primary action, status, and insight
Replaced vague labels with intent-driven copy (e.g., “Invoice Manager” → “Track Payments”)
Removed graphs after feedback showed users didn’t understand them
Used tabs and cards where data was manageable to improve scanability
Brought actions and statuses to the top to guide decision-making

Q3: How do we make data-heavy screens easier to scan?

Studies have found that using colored icons improves visual search speed, recognition, and user satisfaction (Zhang et al., 2022; Microsoft UX Research, 2023).

Most of our screens contain dense operational data, so we used color-coded icons to help users quickly identify task status and data. This not only improves scanability but also aids memory and recognition. All colors are aligned with Delhivery’s design system for visual consistency and accessibility.

Q4: How do we simplify navigation?

We received direct feedback from our Head of Product that even he struggled to understand how the filters worked. That was a wake-up call because if the experience wasn’t intuitive for him, how would it be for less tech-savvy or less formally educated users?

I revisited the filtering experience and took the following steps:

Audited all existing filters across modules
Mapped usage patterns to identify which filters were actually used frequently
Found that the UI was cluttered with quick filters, pop-up filters, and tabs which was causing decision fatigue
Based on this, we prioritized only the most essential filters, placing them upfront and always visible
Aligned filter logic with real-world workflows (e.g., most HRs filter by status or date first)

Q5: How do we keep the flows frictionless for users?

I revisited the existing flows through the lens of progressive disclosure (recommended by NN/g and featured in 100 Things Every Designer Should Know About People.)

The audit revealed that exposing too many data/fields upfront led to cognitive overload and abandonment. By deferring non-essential things and streamlining the interface, we aligned with how people process information: in smaller, goal-oriented steps.

This helped users complete tasks faster, with fewer errors, and no need for external assistance.