Back to blog
Customer Stories

How a D2C Brand Cut Response Time from 4 Hours to 30 Seconds

A Jaipur-based skincare brand was drowning in Instagram DMs and WhatsApp messages. Here's how they automated 85% of customer queries without losing the personal touch.

VS

Varun Sharma

Founder

Feb 3, 20268 min read
How a D2C Brand Cut Response Time from 4 Hours to 30 Seconds

The Problem: Growing Faster Than They Could Reply

NaturGlow (name changed), a direct-to-consumer skincare brand based in Jaipur, was doing well. Too well.

Their Instagram following had crossed 200K. WhatsApp orders were pouring in. But their 3-person customer support team was drowning:

  • Average response time: 4 hours on WhatsApp, 6+ hours on website
  • Missed messages: ~40 per day went unanswered
  • Support hours: Team was working 12-hour shifts
  • Customer complaints: "You never reply" was becoming a common review
  • They needed help — but hiring more agents at ₹25,000/month each wasn't sustainable for a bootstrapped brand.

    The Solution: AI Agent + Human Handoff

    NaturGlow set up an Agent Rush AI agent in February 2026. Here's what they configured:

    Knowledge Base

  • Product catalog (48 SKUs with ingredients, usage, pricing)
  • Shipping and return policy
  • Common skin concern → product mapping
  • Discount and offer details
  • Channels

  • Website chat widget (embedded on Shopify store)
  • WhatsApp Business API
  • Escalation Rules

  • Refund requests over ₹1,000 → human handoff
  • "Allergic reaction" or "skin problem" keywords → immediate handoff
  • Customer asks for human 2+ times → handoff
  • Week 1 Results

    The numbers spoke for themselves:

    MetricBeforeAfter Week 1
    Avg response time4 hours28 seconds
    Messages handled/day120340
    Human agent workload100%15%
    Customer satisfaction3.2/54.6/5
    Missed messages40/day0

    The AI agent handled 85% of conversations end-to-end. The remaining 15% were handed off to the human team — with full context.

    What the AI Handles

    The top conversation types, ranked by volume:

  • "Which product is best for my skin type?" — 31% of queries
  • - AI asks about skin type (oily, dry, combination)

    - Recommends 2-3 products with reasoning

    - Offers to add to cart

  • "Where is my order?" — 24% of queries
  • - AI pulls order status from Shopify

    - Shares tracking link

    - Estimates delivery date

  • "How do I use this product?" — 18% of queries
  • - AI provides usage instructions from knowledge base

    - Shares morning/night routine recommendations

  • "Can I return this?" — 12% of queries
  • - AI explains return policy

    - For eligible returns, generates return label

    - For refunds over ₹1,000, hands off to human

  • "What discounts do you have?" — 8% of queries
  • - AI shares current offers

    - Applies discount codes automatically

    What Gets Handed Off

    The 15% that reaches humans is genuinely complex:

  • Allergic reaction reports (requires medical sensitivity)
  • Bulk/wholesale inquiries
  • Refund negotiations
  • Product complaints requiring investigation
  • Influencer collaboration requests
  • And because the AI passes the full conversation history, the human agent already knows exactly what happened. No "Can you explain your issue again?"

    Month 1: The Financial Impact

    After the first full month:

  • Support cost reduction: ₹50,000/month saved (didn't need to hire 2 additional agents)
  • Revenue increase: 12% lift from faster responses converting to sales
  • Return rate decrease: 8% → 5% (AI recommends better product matches upfront)
  • Team morale: Support team now focuses on VIP customers and complex cases instead of repetitive FAQs
  • What Surprised Them

    "We thought customers would hate talking to a bot. Turns out, they prefer instant answers over waiting 4 hours for a human reply. The ones who do need a human get connected in seconds, with full context. It's the best of both worlds."
    — Founder, NaturGlow

    The biggest surprise was the lead capture. The AI agent captured 340 new leads in the first month from website visitors who would have otherwise bounced. These visitors asked questions, got helpful answers, and left their contact details for follow-up offers.

    Key Takeaways for D2C Brands

  • Start with your top 5 query types — they probably cover 80%+ of volume
  • Set up human handoff from day 1 — it builds trust and catches edge cases
  • Use WhatsApp — it's where Indian customers already are
  • Train on your actual products — generic AI responses don't convert
  • Monitor and improve weekly — review conversations the AI struggled with and add to the knowledge base
  • Try It Yourself

    NaturGlow's setup took 15 minutes. Yours can too.

    Start free at [agent-rush.com](https://agent-rush.com) — no credit card required.

    Share this article
    VS

    Varun Sharma

    Founder

    Building the future of customer support at Agent Rush. Passionate about AI, product design, and creating delightful user experiences.