Hi, I'm Harsh Bhosale · AI Product Designer

I design AI products that respect the people using them.

Two years inside a logistics startup, designing an AI agent system end-to-end across email, WhatsApp, and voice. For a national 3PL with 27 dispatchers across 6 US regions. Now looking for the next senior product design role at a company that takes its audience seriously.

// Open to senior PD roles · 2026 Let's talk →
// Identity
Harsh
Bhosale
AI Product Designer · End-to-end ownership across email, WhatsApp, and voice. 2024–2026 at Debales.
📍 Based in NCR, India · Open to remote
Currently
Open to senior PD roles.
Looking for AI-first product companies. Direct response, no recruiter cycle. Available to start.
// Tenure
2y
2024–2026 · end-to-end at Debales
// Channels
3
Email · WhatsApp · Voice
// Operation
27 / 6
Dispatchers · 6 US regions · 1 client
// Case study
5 DDs
4 published · 1 in progress
// Main case study · 2024–2026
The AI Agent System across email, WhatsApp, and voice. Two years end-to-end. The umbrella story across all five deep dives. Start here.
Read the case study below ↓
// What I believe
I don't trust dashboards that hide their gaps. I don't trust AI products that pretend the seams between AI and human aren't the most important part of the design.
// Read the principles ↓
The story · how I got here · in two years

I joined a small AI logistics startup in late 2024. By the time I left, the AI agent system was running across email, WhatsApp, and voice.

I joined Debales as the team was scaling its first major 3PL relationship. The client was a national logistics operator with 27 dispatchers across 6 US regions, running a delivery operation on email, phone calls, and improvised process. The brief was simple. The work was not.

I started with email. Designed the AI Inbox UI for regional managers and the trust ladder that took the AI from drafting replies in Outlook to sending them autonomously. The audience-separation pattern (managers monitoring through one surface, agents working through Outlook) became the foundation for every channel that came after.

In early 2026 I extended the system to WhatsApp for ETA threads. v1 launched with 0% driver replies. The fix wasn't a redesign. It was format, data, and training, all three at once. That was the moment I learned adoption is rarely a UI problem.

March 2026 brought the Voice AI Agent. Screening only, by design. April 2026 added a softphone for human dispatchers, the last update I shipped before leaving end of April. Phase 3 (resolution) was scoped but not mine to build. I'm now looking for the next senior PD role at a company that takes its audience seriously.

2 years end-to-end 3 channels 5 deep dives 1 client · 1 system Lead designer
01 The work · one case study, five deep dives

Start with the case study. Then dive in.

// 2 years · 3 channels · 1 client
Main case study System · all 3 channels 2024–2026
MCS · The umbrella story

The AI Agent System · email, WhatsApp, and voice for a national 3PL.

"The bottleneck isn't reply speed. It's that the customer's question can't be answered from inside the inbox."

End-to-end design across three channels for a national 3PL with 27 dispatchers across 6 US regions. The umbrella story across all five deep dives. The reframe, the four foundational decisions, the three-channel arc, the six strategic lessons, and the honest scope boundary. Start here.

Read the case study →
3
Channels
email · WhatsApp · voice
27 / 6
Dispatchers · regions
1 client
528 / 95.1%
Voice calls · success
screen + transfer
5 DDs
Deep dives
4 published · 1 in progress
DD/05
Channel · Voice Mar–Apr 2026
Voice AI Agent · screening only, by design.

"The AI screens. The human resolves. The handoff is the design."

528 / 95.1% calls · success
Read DD/05 →
DD/04
Channel · WhatsApp Jan–Mar 2026
WhatsApp Agent · format, data, training, all three at once.

"How 0% became 25% when we pulled three levers, not one."

0% → 25%+ driver reply
Read DD/04 →
DD/02
Channel · Email 2024–2025
Email Dashboard · the system view managers actually used.

"Two surfaces, one system, and a deliberate choice about who sees what."

2 audiences managers · leadership
Read DD/02 →
DD/01
Channel · Email 2024–2025
Email Inbox · audience separation as architecture.

"What started as a feature became the right architecture by the time it shipped."

7 phases inbox arc
Read DD/01 →
DD/03
In progress currently writing
Root Cause Engine · why the agent acts the way it does.

Coming soon. The honest gap in the case study.

In progress honest scope
02 How I think · earned by specific moments

Six principles, none of them slogans.

// every one points back to a deep dive
// Principle 01
Audience separation beats universal design.
Different audiences need different surfaces, formats, and disclosure. Managers see across calls. Drivers tap a button. Customers get a warm reply. Same data, different formats per audience.
Earned in DD/01 · DD/04 · DD/05
// Principle 02
Trust ladders beat big-bang AI launches.
Drafts first. Watch the quality. Move to auto-send only when the team is confident. The same pattern carried from email to WhatsApp without reinvention.
Earned in DD/01 · DD/04
// Principle 03
Honest gaps beat faked numbers.
25% against an 80% target. Sentiment "not available for this call yet." Accuracy 42% inline on a customer turn. The audit trail is only useful if it's true.
Earned in DD/04 · DD/05
// Principle 04
Adoption is a system problem, not a UI problem.
When 0% of drivers replied, the format mattered. The missing 31.2% phone numbers mattered. The training program mattered. Owning the outcome means owning all three levers.
Earned in DD/04
// Principle 05
Pattern reuse beats new architecture per channel.
The two-surface pattern. The trust ladder. The tag system. Same vocabulary across email, WhatsApp, and voice. Less to learn, more pattern coherence, less rebuilding.
Earned in DD/01 · DD/04 · DD/05
// Principle 06
Restraint is a design choice.
Voice v1 didn't try to resolve. WhatsApp v1 only handled ETA threads. Picking a clear scope and executing it well teaches more than half-shipping breadth.
Earned in DD/04 · DD/05
03 Get in touch · direct, no padding

If the work resonates, reach out.

// no recruiter cycle
Open to senior PD roles · 2026

Senior PD roles, conversations about AI product design, or anything that connects to the work above.

I respond to direct outreach, not recruiter cycles. If you've read the deep dives and something resonated, that's the strongest signal I can ask for. I'm based in NCR, India and open to remote.

// Status
Open to roles
Senior product design at AI-first product companies. Available now.
// Location
NCR, India
Open to remote and relocation for the right team.
// Last role
Debales
AI Product Designer · 2024–2026 · end-to-end across three channels.