Routine shared · product links ready
Refund status email · 11:00 PM
Gardening80% AI resolution rate
Order tracking, plant care, replacements, and delivery rescue calls across chat, social, email, and voice.
How Plum handled a 14x Birthday Sale surge with 94% peak AI resolution
How Plum absorbed a 14x Birthday Sale surge with Sagepilot: 101,543 tickets in 30 days, 87% AI resolution, 94% at peak across WhatsApp, email, web, and voice.
94%
peak AI resolution during Birthday Sale
101,543
tickets touched in 30 days
128,463
calls in 30 days
The brand
Plum was built on a clear promise: 100% vegan, cruelty-free beauty that is honest about what goes in the bottle. From green tea skincare to the BodyLovin' range, Plum grew into one of India's most loved clean beauty brands, selling online and on shelves across the country. Every order brings a customer who may ask about a parcel, a refund, a routine, or a skin concern, and beauty customers expect the answer while they are still in the moment.
The challenge
Plum's support queue runs at a scale most consumer brands never reach: 100,000+ tickets and 750,000+ customer messages in a month across WhatsApp, email, and web chat. The patterns repeat relentlessly: shipping and delivery, order and payment, returns and refunds, and thousands of delivery-failure cases where every quiet hour makes a return-to-origin more likely. Then sale weeks arrive. During the Birthday Sale, daily tickets jumped from around 500 to 7,213, the kind of swing no human roster can flex up for and then unwind a week later.
Why Sagepilot
Plum deployed Ananya, an AI employee trained on Plum's catalog, policies, and voice, across WhatsApp, email, the website widget, and phone support. She is connected to Plum's order, logistics, and returns systems, so she does not only explain what is happening. She checks live tracking, schedules reattempts, initiates returns, answers product and ingredient questions, and routes the cases that genuinely need a human with context already attached. Plum's human agents work Indian business hours; Ananya covers nights, weekends, and the gaps between shifts.
The results
In a recent 30-day window, Ananya touched 101,543 tickets and resolved 89,075 end to end, an 87% AI resolution rate. 98% of Plum's replies were AI-generated, and email became one of the fastest channels: 4 in 5 tickets handled by Ananya closed within two replies. The sale surge did not knock the system backward. During the Birthday Sale, AI resolution held between 86% and 94%, with the best days of the month landing mid-surge. Voice now runs as a blended operation too, with 128,463 calls in 30 days split across AI and human agents on one queue.
Peak scale · Birthday Sale
The sale week that held steady
Plum's Birthday Sale pushed daily tickets from around 500 to 7,213. Ananya absorbed the surge across WhatsApp, email, web chat, and connected order workflows, with AI resolution holding between 86% and 94% through the peak fortnight.
Birthday Sale command viewlive peak
~500
usual daily tickets
7,213
peak sale-day tickets
94%
best AI resolution
WhatsApp + web chatabsorbed by AI
Email inboxabsorbed by AI
Delivery-failure queueabsorbed by AI
Post-purchase · Delivery rescue
Failed attempts, caught before returns
Delivery failures are where support becomes revenue protection. Ananya reads the live shipment status, confirms whether the parcel is still in the courier network, and books the reattempt before return-to-origin becomes unavoidable.
Delivery wala aaya hi nahi, app mein attempted dikha raha hai. Order wapas mat bhejna please.
I checked your shipment just now. The courier logged a failed attempt today, but the parcel is still at your local hub. I’ve scheduled a redelivery for tomorrow and added your alternate instructions.
- Live failed-attempt status checked
- Redelivery requested before return-to-origin
8,000+ delivery-failure tickets/month
Pre-purchase · Skin consultations
Advice from the shelf
Plum customers do not only ask where their order is. They ask what goes together, what to start with, and how to build a routine. Ananya answers like a trained beauty advisor, with product context and caution built into the reply.
I have oily skin and dark spots. Should I use niacinamide or vitamin C, or both?
You can use both, but start gently. Niacinamide in the morning can help control oil and marks, and vitamin C can brighten over time. Start with one for two weeks, then layer the second if your skin feels settled.
Oily-skin brightening routine
Niacinamide first · vitamin C next
Routine shared · product links ready
Advice that ends with product links
Coverage · Email
Email that closes in two replies
Refund and order-status emails arrive after hours, but Ananya already has the order context. She verifies the payment and return state before replying, so most AI-handled email threads never need a third message.
Refund status email · 11:00 PMclosed
Where is my refund for order #PLM-2841? It has been a few days.
- Order and return pulled from source systems
- Refund timeline verified before reply
- Complete answer sent in the first response
Voice · Call centre
AI and humans, on one line
Plum's phone support runs as a blended call centre. AI handles the front line and the human team sits beside it for judgment-heavy calls, with handoffs carrying the context instead of making customers start over.
Blended voice operation30 days
72,732
AI-handled calls
55,731
Human-handled calls
2,177h
Total talk time
Inbound calls60,073
Outbound calls68,390
Average call1m 1s
AI and human agents share the same line, queue, and customer context.
Onboarding
Going live
Phase 1 · AI support on chat
Ananya went live on WhatsApp and the website widget, trained on Plum's catalog, support policies, and brand voice.
Phase 2 · Email
The support inbox joined the same brain, turning a slow channel into one where most AI-handled tickets close within two replies.
Phase 3 · Connected actions
Order, logistics, and returns integrations let Ananya resolve, reattempt, and refund instead of just replying.
Phase 4 · Voice
Ananya moved onto the phone line, handling most AI-managed calls while human agents stayed in the same blended operation.
Today · Peak-proof
100,000+ tickets a month, 87% resolved by AI, 128,463 calls in 30 days, and sale weeks that no longer need a war room.
Customers
More stories
Gardening
BeautyRetention that runs itself
Win-back and re-engagement journeys on autopilot.
Fitness80%+ of support automated
Tracking, returns, delivery rescues, and win-back calls across chat, social, and voice.
Join the brands that run their customer experience on Sagepilot
Live in about 48 hours, across every channel your customers use.