Promptly — Ishita UX Case Study

01 · Overview

The process wasn't broken. It was absent.

Resource & Supply's clients place promotional product orders — branded mugs, apparel, corporate swag — but the entire workflow happened over email. A Customer Service Representative would manually pick products, gather brand files, commission proofs, wait for approval, and loop back through revisions. Each cycle added days.

The result: an average turnaround of two weeks per order — not because of production, but because of an ordering process that had never been redesigned for modern expectations.

My role spanned the full stack: I contributed to UX research, led key areas of interaction design, and handled frontend engineering to bring the platform to life.

37%

Projected increase in sales from streamlined ordering

2 wks → 10 min

Order turnaround reduced via self-service proofing

$17.8B

Promotional products industry revenue in 2024

"As a customer placing a promotional product order, I want to ease the ordering process so I can avoid explaining changes over email and reviewing multiple proofs — but ordering through emails causes misunderstandings and delays."

02 · Research & Context

A high-growth market with a process stuck in 2005.

72% of B2B buyers now demand seamless online ordering. 90% prefer customized brand experiences. Yet 4imprint, Vistaprint, BAMKO, and HALO — the dominant players — still lack real-time proofing and interactive customization. The gap was clear.

Our primary persona, John Thompson, is a 40-year-old Sales Manager in Indiana. He orders promotional products regularly for corporate campaigns, is moderately tech-literate, and just wants accurate products delivered on time without the back-and-forth.

📧

Proofing over email broke everything

Proofs sent by email didn't match the sales order, leading to repeated revision cycles that delayed delivery and frustrated clients.

🔁

Repeat proofing caused the biggest delays

The emotional low point in the customer journey was proof review — where mismatched IP caused the entire cycle to restart from scratch.

📋

Limited product discovery

CSRs manually curated a small set of product options, meaning customers rarely saw the full catalog or discovered new items.

📦

Final product often differed from proof

Without live previews, what clients approved and what they received were sometimes meaningfully different — leading to costly dissatisfaction.

03 · Our Solution

Promptly — self-service ordering, built for real people.

We designed Promptly as an AI-powered e-commerce platform that eliminates the CSR from the ordering loop entirely — without losing the quality or accuracy clients need.

🛒 E-Commerce Experience

A familiar, Amazon-style storefront where clients browse the full product catalog, filter by category and price, and place orders without needing a CSR to curate options for them.

"The interface feels intuitive and familiar, almost like shopping on Amazon, which made navigation super easy for me."

👤 User Accounts & Smart Reordering

Persistent customer profiles store past orders and brand preferences, enabling AI-driven reorder recommendations and eliminating repeat setup time for returning clients.

"I like the concept of revisiting my past orders and receiving personalized recommendations based on them."

🤖 AI-Driven Customization

Clients upload brand assets and the AI generates multiple proof variants automatically — starting with popular options, then offering advanced and fully custom AI-suggested designs.

"AI-driven customizations showed me different variations, something I wouldn't have even bothered to check out."

🔍 Live Product Proofs & AR Previews

Real-time AR and 360° views let clients see exactly what their branded product will look like before ordering — eliminating the approval cycle that was costing everyone weeks.

"Seeing a live preview of the product with my logo made it so much easier to finalize my decision."

03.5 · User Journey Walkthrough

From browsing to order confirmed — in under 10 minutes.

Every screen in the flow was designed to eliminate a specific friction point from the old email-based process. Click any screen to expand it.

Step 1 Storefront — browse the full catalog

R&S storefront homepage showing product categories and top sellers

A familiar e-commerce homepage replaces the CSR-curated email. Clients can browse by category and discover products they'd never have known to ask for.

Step 2 Product detail — upload logo, set quantity

Clients drag and drop their logo directly onto a live product preview, pick a color, and set quantity with a price that auto-calculates. No file emails, no back-and-forth on specs.

Step 3 Customization editor — printable area, 3D view, AR

Customization editor showing printable area dimensions, 3D view, and AR preview toolbar

The editor shows exact printable dimensions on a 3D product canvas. Clients can switch between Text, Colors, Images, Shapes, 3D View, and AR Preview — full creative control without needing a design background.

Step 4 AI proof generation — first round of variants

AI proof generation with natural language prompt showing 4 design variants

The client types a natural language prompt and the AI generates four proof variants instantly. A 360° drag view lets them inspect every angle before deciding.

Step 5 AI iteration — refine and regenerate

Second round of AI proof variants showing star pattern designs across the full mug surface

Not happy with the first round? Hit Regenerate. The AI produces an entirely new set of variants. What used to take a week of email rounds now takes seconds.

Step 6 Proof detail — dimensions and final review

Proof detail modal showing selected mug design with exact print dimensions annotated

Before committing, clients can expand any proof to see exact print dimensions annotated directly on the product. This single screen replaces the embellisher proof email — the most frustrating step in the old journey.

Step 7 Order review — quantity, shipping, payment

Order review modal showing product image, quantity, subtotal, shipping address, payment method and arrival date

A clean order modal surfaces everything at once — quantity, live price, shipping address, payment method, and estimated arrival. No surprises, no separate confirmation emails.

Step 8 Order confirmed — done

Order confirmed modal showing final product image, order summary and continue shopping button

Order placed. No follow-up email needed, no CSR loop to close. The client has a confirmed order, a delivery date, and an invoice — all self-served in under 10 minutes.

04 · Design & Engineering Process

From client journey maps to a working frontend.

1

Client Journey Mapping

We mapped the full ordering journey across six touchpoints. The map revealed the emotional low point was proof review from the embellisher — where mismatch with approved IP caused frustration and delays. This became the primary design target.

2

Competitive Analysis

We benchmarked 4imprint, Vistaprint, BAMKO, and HALO. All four excel at product variety but none offer real-time proofing, live AR previews, or AI-driven design suggestions. We identified this as Promptly's core differentiator.

3

Persona & Tiered Design Framework

We designed around three customer types: the time-pressed buyer, the hands-on customizer, and the creative power user. This tiered approach ensured no user felt overwhelmed while still offering depth for those who wanted it.

4

IoT Architecture & Technical Design

I helped design the full system architecture: mobile camera sensors for AR spatial data, internet & mobile network transmission via HTTPS/REST APIs/WebSockets, AWS cloud for product catalogs, AI model hosting, and security management.

5

Frontend Engineering

I led frontend development, building the e-commerce storefront, product customization interface, and proof preview components. The goal was a familiar, fast, and intuitive experience that required no training for end users.

6

Prototype & Pitch

We packaged the full solution — market research, user research, business model canvas, SWOT analysis, and revenue model — into a pitch delivered to Resource & Supply stakeholders.

05 · Technical Architecture

A full-stack system designed for scale.

Layer	Technology	Purpose
Sensors	Mobile Phone Camera	Captures spatial data for real-world AR product previews
Communication	Internet & Mobile Network	Transmits data between client device and AWS backend
Protocols	HTTPS · REST APIs · WebSockets	Secure, real-time data transfer and proof updates
Cloud	AWS	Hosts product catalogs, user data, generated proofs, AI models, and security
Catalog	ASI's API + SAGE / ESP / CommonSKU	Sources promotional product inventory from distributor networks
Frontend	React · Shopify / CommonSKU	Self-service e-commerce storefront with embedded AI features

06 · Business Model

Two revenue streams. One compounding advantage.

Promptly creates value in two directions: directly for R&S clients through the e-commerce platform, and horizontally for other distributors through an AI proof generation API.

Primary — E-Commerce Sales

The Shopify or CommonSKU storefront generates revenue from direct promotional product sales. AI personalization improves conversion rates and increases average order value as users discover new products and variations.

Secondary — SaaS Subscription (API Access)

Other promotional product distributors can license Promptly's AI proof generation tool via a tiered monthly subscription:

$50

/month

50 Proofs

$150

/month

500 Proofs

$250

/month

2,000 Proofs

$500

/month

Unlimited Proofs

07 · Validation

What the pitch revealed — and what we'd do next.

❌ Current State Problems

Proofs arrive via email, hard to annotate or track changes
Limited product options surfaced to clients by CSRs
No live preview — final product often differs from proof
CSRs stuck doing manual, repetitive order management
2-week average turnaround due to back-and-forth cycles

✅ Promptly's Solutions

Collaborative real-time proofing with inline commenting
Full self-browse catalog with filters and reorder recommendations
Live AR/360° preview before any order is placed
CSRs freed to focus on lead generation and sales
Projected turnaround reduced from 2 weeks to 10 minutes

"The self-service model doesn't remove the human from the process — it removes the friction. CSRs get their time back. Clients get their autonomy back."

08 · Outcomes & Reflection

What I learned about designing for real business constraints.

Research → Engineering

Being involved in both research and frontend development meant I understood the "why" behind every UI decision. That context made my engineering choices better — not just technically, but experientially.

B2B ≠ boring UX

B2B users have the same expectations as consumers. John Thompson doesn't want a clunky portal — he wants Amazon. Designing for that insight changed how I approached every screen.

Systems thinking matters

Designing the IoT ecosystem taught me to think beyond the interface — how data flows, where latency affects UX, and how backend decisions shape what's possible on the frontend.

What I'd do next

Usability testing with actual R&S clients, A/B testing the proof generation flow, and measuring real order turnaround times post-launch to validate the 37% sales projection against actual data.