Presenter Notes
Prepared for Gaurangi Dangat · Salasar Auction
AI Voice Agents that call leads, run auctions, and never miss a follow-up.
Two AI calling systems — built and run for Salasar Auction. Meta lead follow-up. Auction notification and reminders. CRM-connected. Hindi-first.
AI Voice Agent Outbound Calling CRM Integration Hindi + English Auction Automation
ConverseAI — theconverseai.com July 2026  ·  Confidential
About ConverseAI

We build AI calling systems — and keep them running.

AI Voice Agent
WhatsApp AI
CRM Integration
Live Chat AI
Analytics
50+
Businesses Served
100+
AI Systems Built & Running
300+
Integrations Delivered
5+
Years Building & Running AI
Meta Tech Provider Partner  ·  GDPR & CCPA Aligned  ·  Founded 2021  ·  We don't train on your data
The problem
Manual calling at scale
means lost leads and missed auctions.
Two completely different calling jobs. Both currently dependent on humans dialing one contact at a time — creating delays, gaps, and inconsistent outcomes.
✕

Meta leads go cold before anyone calls

Leads generated via Meta Ads have a very short attention window. Manual calling teams can't respond fast enough — intent drops within minutes of the form submit.

✕

Auction calls are a manual slog

Before every auction, someone has to call each relevant buyer — filtering from the database, dialing individually, logging outcomes. It's slow, inconsistent, and doesn't scale.

✕

Reminders are missed or skipped

Following up with interested buyers before the auction date requires another round of manual calling. In practice, many reminders don't happen — participation suffers.

✕

CRM is never fully updated

Call outcomes, interest levels, and stage changes depend on reps manually updating records. Data is always incomplete — reporting and follow-up decisions suffer.

✕

Rigid scripts kill the conversation

Any AI or IVR solution tested so far used fixed flows. The moment a buyer asked an unscripted question, the system broke — lowering trust and conversion.

✕

Scaling means more headcount

As auction volume grows, so does the calling workload. The current model means every new auction adds more manual burden — there's no leverage.

The solution
Two AI Voice Agents.
One for leads. One for auctions.
Use Case 1 · Meta Ad Leads

Calls new leads the moment they come in — qualifies interest and moves them to subscription.

Meta lead arrives → AI agent calls within minutes → understands intent → progresses or disqualifies → CRM updated with outcome.

  • Calls leads immediately — no queue, no delay
  • Natural conversation — not a script reader
  • Qualifies interest in subscription platform
  • Captures outcome and updates CRM automatically
Use Case 2 · Auction Calling

Pre-auction outreach and reminders — automated, stage-aware, and CRM-synced.

CRM filters relevant buyers → AI agent calls to inform about upcoming auction → logs confirmed interest → sends reminders closer to auction date.

  • Identifies relevant buyers from CRM by category
  • Pre-auction notification calls at scale
  • Reminder calls for confirmed and interested buyers
  • Stage-wise calling based on buyer status
Use Case 1 — Meta Ad Lead Follow-Up

Call every lead.
Within minutes.
No follow-up missed.

  • ⚡
    Instant outreachTriggers the moment a Meta lead is submitted — no waiting for a rep to be free.
  • 🧠
    Dynamic conversationListens to what the lead says and responds in context — not a fixed script.
  • 🎯
    Intent qualificationIdentifies genuine interest in the auction subscription and flags it for the sales team.
  • 📋
    Auto CRM updateCall outcome, interest level, and next step written back to CRM — zero manual entry.
AI Agent · Hindi/English
"Hello! Main Salasar Auction ki taraf se bol raha hoon. Aapne recently humare auction platform ke baare mein enquiry ki thi — kya aap 2 minute mein baat kar sakte hain?"
Lead
"Haan bolo, main steel scrap mein deal karta hoon."
AI Agent
"Bilkul! Hamare platform pe steel aur ferrous scrap ke auctions regularly hote hain. Aap subscription leke har auction ki notification seedha apne phone pe pa sakte hain — aur directly bid bhi kar sakte hain. Kya aap aur jaanna chahenge?"
Lead
"Haan, charges kya hain?"
Intent confirmed → Transferring to sales team with full summary
Auction Calling Flow
1
Auction is created in CRM
Material category, auction date, lot details entered by Salasar team
2
AI identifies relevant buyers
Queries CRM for buyers registered in that material/scrap category
3
Pre-auction notification calls
AI calls each buyer, informs about upcoming lot, gauges interest, answers questions
4
Interest captured → CRM updated
Confirmed / interested / not-interested logged per buyer automatically
5
Reminder calls before auction
AI auto-calls confirmed + interested buyers 24–48 hrs before the auction date
6
Participation tracked → CRM closed
Post-auction status updated; complete calling history retained per buyer
Use Case 2 — Auction Calling

Every relevant buyer
called, notified,
and reminded.

The AI reads your CRM, identifies which buyers need to know about an upcoming auction, calls them in natural conversation, and keeps track of every response — so your team focuses on the auction, not the calling.

  • 🏭
    Category-aware outreachOnly contacts buyers registered for the relevant material/scrap type.
  • 🔔
    Staged callingNotification → interest confirmation → reminder — each stage automated.
  • 📊
    Full CRM write-backEvery call outcome stored — complete visibility into buyer engagement.
How it works
From data to conversation
to CRM — end to end.
AI handles the calling. Your team handles only what the AI can't.
1

Trigger received

New Meta lead or auction date from CRM kicks off the calling flow

2

AI places the call

Calls the contact in Hindi or English — greets naturally, states the purpose

3

Dynamic conversation

Listens, responds in context, handles objections, answers questions

4

Outcome captured

Confirmed / interested / callback / not-interested — classified in real time

5

CRM updated

Full call summary, outcome, and next step written back — zero manual entry

CRM Integration
Your CRM stays the
system of record.
The AI agents don't replace your CRM — they work inside it. Every call decision is driven by your data. Every call outcome is written back in.
📥

What the AI reads from CRM

  • Buyer/customer database
  • Material/scrap category registrations
  • Auction calendar and lot details
  • Customer stage and calling status
  • Meta lead records (on new enquiry)
📤

What the AI writes to CRM

  • Call outcome per contact
  • Interest level (confirmed / interested / not-interested)
  • Call transcript summary
  • Next action or follow-up date
  • Escalation flag (if human handover needed)
🔗

Integration approach

  • API or webhook-based connection to your CRM
  • Real-time bidirectional sync on call events
  • Configurable field mapping to match your data structure
  • Works with most major CRMs (Zoho, Leadsquared, custom)
❓

What we need to confirm

  • Which CRM is currently in use
  • API availability and authentication method
  • Current data structure and key field names
  • Read/write permissions and IT contact
CRM details will shape the final technical architecture. A 30-minute session with your tech/ops team is enough to map this completely.
Conversation quality
Not an IVR. Not a script reader.
A real conversation.
What you experienced before

Rigid, scripted voice AI

  • Fixed decision tree — every response had to match a preset path
  • Caller deviates → system gets stuck or repeats the same line
  • Robotic voice quality — customers recognize it immediately
  • No real understanding — pattern matching, not comprehension
  • Any unscripted question leads to dead air or escalation
ConverseAI approach

Dynamic, reasoning-based AI

  • Understands what was actually said — intent, not keywords
  • Responds in context even when the conversation goes off-path
  • Natural Hindi, English, and Hinglish — adapts to the caller's language
  • Handles objections, questions, and hesitation naturally
  • Knows when to escalate — and prepares a full handover summary
Voice quality and conversation quality are evaluated in the pilot — so you can hear and judge it directly before committing to full scale.
Languages
Hindi-first. English-ready.
Regional languages on roadmap.
Pilot launches in Hindi and English. Tamil, Kannada, and other regional languages added in a subsequent phase as your buyer geography expands.
हिंदी
Hindi
Pilot — Primary
English
English
Pilot — Required
தமிழ்
Tamil
Phase 2
ಕನ್ನಡ
Kannada
Phase 2
मराठी
Marathi
తెలుగు
Telugu
ગુજરાતી
Gujarati
বাংলা
Bengali
The agent automatically detects and matches the caller's language — no prompt or menu needed. Mid-conversation language switching (Hindi ↔ English / Hinglish) is handled natively. Final language list for Phase 2 to be confirmed with Salasar before phase design.
How we work
We build it. We train it.
We run it. You see results.
Dialler platform

Tool + subscription

  • You get a calling platform
  • You configure the conversation flows yourself
  • You manage quality and retrain when it fails
  • Support is a ticket queue
Dev vendor

Built, then handed over

  • Custom system built to your spec
  • Delivered and signed off
  • Your team operates it from day one
  • Fixes and improvements = new scope & cost
ConverseAI ✓

Managed AI as a service

  • We scope, build, and deploy the agents
  • We train them on your buyer language and auction flows
  • We monitor performance and improve monthly
  • You see outcomes — not a platform or codebase to manage
Pilot Proposal
Start with one flow.
Validate. Then scale.
Phase 1 — Pilot

One use case. Fully built and running.

  • Select one calling flow — recommend Meta Ad Lead follow-up as the starting point
  • Build conversation design for the selected flow in Hindi + English
  • Connect to your CRM for lead data read and outcome write-back
  • Set up telephony, deploy agent, test with real calls
  • Monitor conversation quality, CRM accuracy, and call outcomes
  • Refine based on live data — we run the iterations
Phase 2 — Scale

Add the second use case + regional languages.

  • Add Auction Calling (Use Case 2) with full stage-wise flow
  • Build pre-auction notification, interest confirmation, and reminder flows
  • Expand to Tamil, Kannada, or other regional languages as needed
  • Full CRM bidirectional sync for both calling systems
  • Ongoing monitoring, performance reporting, and monthly improvements

Commercial Structure

1
One-Time Setup & Development
Covers conversation design, AI agent build, CRM integration, testing, and deployment for the pilot use case.
2
Usage Cost — Per Minute
Recurring cost based on actual AI calling minutes consumed. Includes telephony, AI processing, and platform costs.
3
Monthly Managed Retainer
Ongoing monitoring, conversation quality maintenance, CRM integration support, and improvements.
Precise numbers will be shared after a 30-minute scoping session — we need call volume estimates and CRM details to quote accurately.
Next step

Ready to scope the
right pilot for Salasar Auction.

A 30-minute scoping session — with your ops and CRM team — is enough for us to finalize the technical architecture and give you a concrete proposal with timeline and pricing within a week.

✉ himanshu@revtidigital.com
✉ contact@theconverseai.com
📱 +91 99823 23333
🌐 theconverseai.com