An outbound voice AI campaign is only as good as its configuration. The best AI model in the world delivers poor results if calls are going out at the wrong time, in the wrong language, without proper retry logic. This guide walks through every step of setting up a production-ready outbound campaign in Agni.
Step 1: Prepare Your Contact List
Agni accepts contact lists via CSV upload or API. The minimum required fields are:
- phone_number: 10-digit Indian mobile number (no country code needed; Agni adds +91 automatically)
- name: Customer name (used for personalization in the call script)
Optional but recommended fields:
- language: Preferred language code (hi, ta, te, kn, mr, bn, gu, pa, ml, en). If omitted, Agni auto-detects from first utterance.
- account_id: Your internal customer/account IDused for CRM sync and memory lookup
- custom_1 through custom_5: Any custom variables referenced in your script (EMI amount, due date, product name, etc.)
DNC scrubbing: Before upload, Agni automatically checks your list against your account's DNC registrynumbers flagged as do-not-call are suppressed before the campaign begins. No manual scrubbing required.
Step 2: Configure Your AI Agent
If you haven't already created an agent for this campaign, the Agent Builder in the Agni dashboard lets you configure:
- Script/persona: The AI's opening line, conversation flow, and goal (collect payment commitment, book appointment, qualify lead)
- Voice profile: Male/female, language, dialect, emotion profile (warm/neutral/firm)
- Handling rules: What to do if the customer says no, asks for a callback, or uses specific phrases
- Transfer rules: When and where to transfer to a human agent
For standard use cases (collections reminder, sales lead qualification, appointment reminder), Agni provides pre-built agent templates that can be customized in under 30 minutes.
Step 3: Configure Compliance Settings
This step is non-negotiable for any Indian business. Agni's compliance configuration panel sets:
Call Window
Default: 8:00 AM – 7:00 PM IST. This window is enforced at the infrastructure levelcampaigns cannot be started outside this window, and calls scheduled to cross the 7 PM boundary are automatically held until 8 AM the next day.
Maximum Daily Contact Attempts per Number
Default: 2 attempts per number per day. Configurable up to 3 (RBI market practice). Once the daily limit is hit, the number is held until the next call window opens.
Maximum Total Campaign Attempts
Configure the total number of contact attempts across the campaign lifecycle (e.g., 6 total attempts over 3 days before marking as non-contactable).
DPDP Consent Disclosure
Select your disclosure template. Every call opens with the configured disclosure text in the customer's detected language. Consent acknowledgment is recorded automatically.
Step 4: Configure Retry Logic
Retry logic determines what happens when a call doesn't reach a human:
| Outcome | Default Retry Behavior | Configurable? |
|---|---|---|
| No answer | Retry after 2 hours in same call window | Yes (delay, max retries) |
| Voicemail / AMD detected | Retry next day at different time | Yes |
| Busy signal | Retry after 30 minutes | Yes |
| Switched off | Retry next day | Yes |
| DNC requested by customer | No retrypermanent DNC flag | No (compliance-locked) |
| Callback requested | Schedule for requested time | N/Acustomer-specified |
Intelligent retry scheduling also considers time-of-day patterns: if a customer consistently doesn't answer in the morning, Agni shifts their retry window to afternoon automatically after 2 failed morning attempts.
Step 5: Schedule and Launch
Once configured, set your campaign start time and launch. You can schedule campaigns to start immediately or at a specific future date/time. For large campaigns (100,000+ records), Agni's infrastructure scales automaticallyno prior notice or capacity planning needed.
Soft Launch Recommendation
For new campaigns or new scripts, we recommend a soft launch: start with 5–10% of your contact list, monitor the first 500–1,000 calls in the live analytics dashboard, review a sample of call recordings, and make any script adjustments before launching the full campaign. This saves significant rework if the script needs tuning.
Step 6: Monitor in Real Time
The Agni campaign dashboard shows live metrics during active campaigns:
- Calls dialed / answered / completed in real time
- Outcome distribution (promise, callback, no answer, DNC, transfer)
- Sentiment distribution across completed calls
- Language distributionwhat languages are your customers actually speaking?
- Live call logindividual call outcomes as they complete
"We ran our first 5,000-call campaign on Agni with zero ops overhead on launch day. The dashboard gave us everything we needed to monitor in real time from a single screen."Growth Lead, D2C Brand (Gurugram)
Step 7: Post-Campaign Analysis
After campaign completion, Agni generates a campaign summary report: total contacts reached, outcome breakdown, sentiment summary, language breakdown, and cost summary. Export to CSV or sync to your BI tool via webhook.
Ready to get started?
Launch your first outbound campaign today at app.ravan.ai. Full campaign configuration documentation at docs.ravan.ai. Questions? Write to info@ravan.ai.
