Use Cases
Objective:
Develop an AI voice agent that automates customer verification, EMI reminders, payment collection, and payment plans and remind customer about outstanding payment, record their response (Paid/Promise to Pay/Dispute) and escalate if needed.
Step-By-Step Use Case:
Scenario 1. Customer Verification
- Trigger Event: The AI voice agent calls the customer, who has taken out a consumer loan.
-
Action:
- The AI asks, "Hello, this is [Company Name]. Are we speaking with [Customer's Full Name]?"
- If the customer confirms the identity, the conversation continues.
- If the customer does not confirm, the agent will ask for a specific piece of information (e.g., date of birth or last four digits of their social security number or loan number) to verify their identity.
- Outcome:
- If the customer is verified successfully, the agent proceeds to the next step.
- If verification fails, the AI agent will prompt the user with a polite message and end the call with a request to contact customer service.
Scenario 2. EMI Reminder
- Trigger Event: After successful verification, the AI agent proceeds with a reminder call.
-
Action:
- The AI says: “I am calling to remind you about your upcoming EMI payment for your loan with [Company Name]. Your next EMI is due on [Due Date]. The amount is [Amount].”
- It will also mention any other details relevant to the payment, such as the loan term or previous payments made.
- Outcome:
- The customer is informed about the due EMI and its due date.
Scenario 3. Payment Collection
- Trigger Event:Trigger Event: After the reminder, the AI agent asks the customer if they are ready to make a payment.
-
Action:
- If the customer says, “Yes,” the AI will ask for the payment method, e.g., credit card, debit card, or other available methods.
- The agent will then share a secure payment link over the phone or text message, depending on what communication method is available.
- The AI will say, “To make the payment now, please follow this link: [Link].”
- Outcome:
- Customer is provided with a payment option and a direct link to pay.
Scenario 4. Payment at a Later Date (Post-Due Date)
- Trigger Event: If the customer wants to pay after the due date or requests a later date to pay..
-
Action:
- The AI will notify the customer of the following: “Please be aware that if you pay after the due date, there will be a late fee of [Late Fee] and interest charges of [Interest Charges] on your outstanding balance.”
- The AI will then provide the customer with a payment link and offer to set up a plan if the customer is unable to pay the full amount.
- Outcome:
- Customer is informed about the consequences of paying after the due date, including late fees and interest charges.
Scenario 5. Payment Plan Setup
- Trigger Event: If the customer requests a payment plan, either because they cannot pay in full or they have promised to pay at a later date.
-
Action:
- The AI asks: “Would you like to set up a payment plan to help with your outstanding balance?”
- If the customer agrees, the AI will gather details like how much they can pay monthly or at a specific date.
- The AI will confirm the payment plan: “Your payment plan will be set at [Amount] per month starting from [Start Date], and the final payment will be on [End Date].”
- The AI will ask for confirmation from the customer.
- Outcome:
- The payment plan is set up, and the customer is informed of the agreed-upon schedule.
Scenario 6. Reminders for Payment Plan
- Trigger Event:The payment plan is active, and future reminders are necessary.
-
Action:
- The AI will call the customer ahead of the due date for each installment, saying: “This is a reminder that your payment of [Amount] is due on [Date].
- The AI will also mention any outstanding balances and notify the customer if there are any issues with the payment plan.
- Outcome:
- The customer receives timely reminders for payments.
Scenario 7. Handling Payment Issues
- Trigger Event: If a payment fails, or the customer requests assistance in case of financial issues.
-
Action:
- The AI will inform the customer of the failed transaction and ask: “Your last payment of [Amount] did not go through. Would you like assistance with making the payment again or rescheduling it?”
- If needed, the AI may offer to set up a revised payment plan or connect the customer with a human representative if more support is required.
- Outcome:
- The customer receives support in resolving payment issues, either through rescheduling or by offering alternatives.
Scenario 8. Payment Confirmation
- Trigger Event:The customer successfully makes a payment.
-
Action:
- The AI will confirm: “Thank you for your payment of [Amount]. Your loan balance is now [Remaining Balance], and your next due date is [Next Due Date].”
- The agent will also confirm any late fees or interest charges, if applicable.
- Outcome:
- Customer receives confirmation and updated balance details.
Scenario 9. Ending the Call
- Trigger Event:The transaction is complete, or the payment plan is set.
-
Action:
- The AI will wrap up the conversation by saying: “Thank you for your time. If you have any questions or need further assistance, please don’t hesitate to contact us at [Customer Service Number/Website].”
- The call will be ended professionally and politely.
- Outcome:
- The call concludes with all relevant information shared, and the customer is left with the necessary resources for follow-up.
Additional Features & Considerations:
- Multi-Language Support: The AI voice agent should be capable of handling multiple languages, based on customer preferences.
- Security: All sensitive information (payment details, verification information) should be handled securely with encryption and PCI compliance.
- Escalation Protocol: If at any point, the customer requests to speak to a human agent, the AI should be able to escalate the call smoothly.
- Automated Reporting: The AI should generate a log of all interactions for internal use, including customer verification details, payments made, payment plans set, and issues encountered.
- Error Handling: If the AI cannot understand the customer, it should politely ask for clarification or offer alternative ways to resolve issues
Reward Criteria for the Hackathon:
- Functionality: The AI agent should handle all outlined tasks efficiently, including verification, reminders, payment processing, and late fee management.
- User Experience: The agent should be easy to interact with, providing clear instructions and options to the customer.
- Security and Privacy: Ensuring that customer data and transactions are securely handled.
- Innovation: Creative ways of improving the process, like using NLP techniques for better conversation flow or optimising payment processing steps.
Objective:
Build voice-based AI agents that can assist users with searching real estate properties and scheduling appointments with agents, call back later if requested and propose site visits. simulate data using MagicBrick/ 99Acre APIs etc.
Scenario 1: Initial Property Inquiry Call
The AI agent calls the user (or answers an incoming call) and asks:
- What type of property are you looking for? (e.g., apartment, villa, commercial space)
- Preferred location?
- Budget range?
- Any specific requirements? (e.g., number of bedrooms, parking, pet-friendly)
Scenario 2: Suggesting Matching Properties
The AI agent calls the user (or answers an incoming call) and asks:
- Shortlists and suggests 2-3 matching properties
- Shares brief details on each (location, price, size, features)
- Asks which one(s) the user would like more information about or visit
Scenario 3: Appointment Scheduling
Once the user shows interest:
- Once the user shows interest:
- Confirms a date & time that works for the user
- Optionally books the calendar and sends confirmation via SMS/email
Scenario 4: Appointment Reminder and Reconfirmation (Day Before or Same Day)
- Calls the user to remind them of the upcoming appointment
- Asks if they are still available or if they want to reschedule
- Updates the schedule accordingly and confirms again
Scenario 5: Follow-Up Call After Site Visit
The AI follows up:
- Asks for feedback on the property visit
- Checks if the user is interested in proceeding or would like to explore other options
- BuChecks if the user is interested in proceeding or would like to explore other optionsdg
Scenario 6 : Handling Missed Calls or No Shows In case of a missed appointment:
- AI calls to apologise and reschedule
- Offers alternate timings
- Optionally asks for the reason to improve future planning
Scenario 7 : Handling Cold Leads / Lead Revival For older leads that went cold:
- AI reaches out and asks if the user is still interested in buying/renting
- Offers new properties that match market trends or deals
- Reinitiates the appointment funnel
Objective:
AI Voice agent must call leads, confirm basic info (e.g., name, service interest), propose available slots, book appointment in system
Scenario 1: Cold Outreach Introduction
Objective:Test basic conversation skills and clarity in introducing the product.
Context:The AI agent is initiating a conversation with a prospect who has never heard of the SaaS product.
Expected Capabilities:- Brief, engaging self-introduction
- Clear value proposition
- Ask relevant qualifying questions (e.g., role, industry, needs)
- Capture lead’s name and email
Scenario 2: Website Chatbot for Product Discovery
Objective:Simulate real-time web chat experience for inbound interest.
Context:A visitor lands on the website. The AI initiates a conversation to guide them.
Expected Capabilities:- Greet visitor and identify their goals
- Recommend relevant product features
- Offer demo scheduling or resource download
- Handle FAQs (pricing, integrations, support)
Scenario 3: Objection Handling
Objective:Test AI’s ability to maintain engagement and counter hesitations.
Context:The AI agent is initiating a conversation with a prospect who has never heard of the SaaS product.
Expected Capabilities:- Empathetic response to objections
- Clarify misunderstandings
- Offer comparative advantages or testimonials
- Guide toward a CTA (e.g., free trial, demo)
Scenario 4: Qualification Conversation
Objective:Qualify the lead based on specific criteria (BANT, CHAMP, etc.).
Context:The prospect shows interest, and the agent must gather data for sales team handoff.
Expected Capabilities:- Ask strategic questions: budget, authority, timeline, etc.
- Categorise lead as hot/warm/cold
- ORecord lead details in CRM-ready format
- Ask permission to pass details to human sales rep.
Scenario 5: Follow-up & Nurture Flow
Objective:Show continuity and memory retention across multiple sessions.
Context:The prospect has interacted before but didn’t convert. AI needs to re-engage.
Expected Capabilities:- Recall past interaction
- Provide new offer (e.g., webinar, case study)
- Use urgency (limited-time discount, feature update)
- Escalate to human if qualified
