← All case studies

    Case Study

    Heterize: how Amazon seller onboarding scaled with Hinglish voice AI

    What 400,000 AI calls to Amazon sellers in India taught us about finding the sellers who are ready to start, and how OmniDimension's voice AI qualified them in under two minutes.

    Published November 16, 2026

    Heterize Infotech logoOmniDimension logo

    400,000+

    AI calls to Amazon sellers across India

    200,000

    Sellers reached

    12,000+

    Interested sellers qualified as hot leads

    5,500+

    Hot leads with a regular GST number, ready to onboard

    How a Heterize seller call works: a seller list becomes a campaign, the Hinglish AI agent finds the problem and checks eligibility, hot leads go to the onboarding team and busy sellers go to a callback queue
    Click or tap the diagram to open it full size.

    At a glance

    Customer
    Heterize Infotech, an Amazon seller onboarding and growth partner in India
    Industry
    E-commerce services
    Goal
    Find registered and new sellers who want to start selling on Amazon, and route the ready ones to the onboarding team
    Channel
    Outbound AI calls in Hinglish
    Scale
    10+ campaigns over 13 months, each with its own pitch

    Overview

    Heterize Infotech helps sellers in India get started and grow on Amazon. Its lists hold hundreds of thousands of sellers: some registered but not yet live, some new, and some selling with room to grow. Each seller needs a short, personal conversation before the onboarding team can help.

    OmniDimension gave Heterize an AI agent that calls each seller in natural Hinglish, finds the problem, pitches Amazon's onboarding support, checks eligibility and sends only the ready sellers to the onboarding team.

    The challenge

    The problem

    A seller list is long, and most sellers are not ready today. The onboarding team needed answers to three questions on every call:

    1. Is this seller interested in selling on Amazon now?
    2. What is stopping them? Listings, catalogue, account setup or plain doubt.
    3. Are they eligible? A regular GST number, a trademark or brand authorization, and a product category.

    Why calling by hand did not scale

    • A human team can make a few hundred calls a day. The lists held hundreds of thousands of sellers.
    • Most calls end in under half a minute, so callers spent their day on people who were not ready.
    • Sellers across India answer in Hindi, English or a mix, and a formal script sounds wrong in all three.
    • Eligibility checks happened late, so the onboarding team spent time on sellers who could not go live.

    Key insights from 400,000 seller calls

    1. Sellers who have registered convert best

    Lists of sellers who had registered but not started gave twice the share of onboarding-ready leads. About half of their hot leads already had a regular GST number, against about a quarter on general seller lists. The best list is the one where the seller has already taken the first step.

    2. A no takes 30 seconds, a yes gets two minutes

    A typical connected call lasted under 30 seconds. A call that ended in a hot lead lasted about 100 seconds. The agent lets uninterested sellers go fast and spends its time on the ones who lean in. That is how 400,000 calls stay affordable.

    3. "Call me later" is a lead, not a no

    More than 10,000 sellers asked for a callback at a better time. The agent recorded the date and time on each one, so a busy seller became a scheduled call instead of a lost one.

    4. Hinglish wins the conversation

    The agent speaks the way sellers talk: Hindi and English mixed, with everyday English words like "seller", "listing" and "GST". Sellers answer in both languages, and the agent captures their answers either way, for example a product category said in Hindi.

    5. Ask eligibility before the hand-off, not after

    Every interested seller was asked about GST, trademark or brand authorization, and product category on the same call. The onboarding team got a short list of sellers who could go live, with the facts already filled in.

    The solution: a five-step qualification engine

    1. Problem discovery

    The agent opens with one direct question: is the seller facing an issue with listings or account support? This finds the reason the seller has not started, in the seller's own words.

    2. Benefits pitch

    The agent explains the offer: three months of free account management and a dedicated account manager for listings, catalogue and sales support.

    3. Eligibility screening

    For interested sellers, the agent checks:

    • a regular, active GST number (composite GST does not qualify)
    • a trademark, brand authorization or purchase invoice
    • the product category

    4. Objection handling

    If the seller hesitates, the agent asks what is holding them back, answers the concern, and offers a one-minute eligibility check. If the seller is busy, it books a callback time.

    5. Smart routing

    Each call ends with structured data: interest, hot lead status, GST, brand documents, category and callback time. Hot leads go to the onboarding team, and callbacks go back into the calling queue.

    What made it work

    • Natural Hinglish with simple words, never formal translated Hindi.
    • One agent per campaign, so each list got the pitch that fit it.
    • Plug and play: new lists went live as new campaigns without changes to the team's process.

    How it works

    The diagram at the top of this page shows the flow:

    1. Heterize uploads a seller list as a campaign.
    2. The agent calls each seller and asks about their Amazon problem.
    3. Interested sellers hear the benefits and get an eligibility check.
    4. Hesitant sellers get their concern answered. Busy sellers get a callback time.
    5. Every call writes back interest, eligibility and next step.
    6. Hot leads go to the onboarding team. Callbacks go back into the queue.

    The impact

    400,000+
    AI calls to Amazon sellers
    About 200,000
    Sellers reached
    About 150,000
    Connected calls
    12,000+
    Interested sellers qualified as hot leads
    5,500+
    Hot leads with a regular GST number, ready to onboard
    10,000+
    Callback times captured from busy sellers
    Under 2 minutes
    Typical length of a call that ends in a hot lead

    The onboarding team works only on ready sellers. About 1 in 12 connected calls became a hot lead, and each one came with eligibility already checked.

    The reach no human team could match. 200,000 sellers got a personal call in their own mix of Hindi and English.

    Less manual calling. The team stopped dialling lists and started onboarding sellers.

    The outcome

    Heterize runs seller onboarding as a steady pipeline. The AI agent makes the first call, finds the problem, makes the pitch, checks eligibility and books callbacks. The onboarding team starts every day with a list of interested, eligible sellers instead of a raw phone list.

    Lessons for e-commerce enablers and marketplaces

    1. Call the sellers who already started. Registered sellers convert to ready leads at twice the rate.
    2. Let a no end fast. Short calls with uninterested people keep large campaigns affordable.
    3. Treat every callback as a lead. Record the time and call back.
    4. Speak the seller's language. Natural Hinglish beats a formal script.
    5. Qualify before the hand-off. Ask the eligibility questions on the first call.

    See more: voice AI for e-commerce, multilingual voice AI, voice AI for Indian languages, and OmniDimension for e-commerce.

    FAQs

    See what OmniDimension can do for your leads

    Book a 15-minute walkthrough with our team.

    Book a Demo