We are planning to use Retell AI for a large-scale outbound debt collection calling campaign and wanted to confirm a few things before we finalize our implementation.
Our Use Case: We need to make 10,000 outbound AI calls daily for invoice follow-up and debt collection using a Retell AI agent.
Our Planned Approach: We plan to use the Batch Call API (POST /create-batch-call) to send calls in groups of 50 tasks per API call. So instead of making 10,000 individual API calls, we will make 200 API calls with 50 tasks each — totaling 10,000 calls per day.
Questions we want to confirm:
Is this approach correct and supported? Can we send 50 tasks per batch API call and make 200 such API calls to cover 10,000 calls?
How many phone numbers do we need to purchase for this volume? Can we use 1 from_number for all 10,000 calls, or do we need multiple numbers?
What is the maximum number of concurrent calls supported on the Enterprise plan? We are targeting at least 50-100 concurrent calls running simultaneously.
Approximately how many hours will it take to complete 10,000 calls with 50-100 concurrent calls?
Is there any rate limiting on the Batch Call API that we should be aware of?
What Enterprise plan pricing would you recommend for this volume?
Please advise on the best setup for our requirements.
Quick question regarding concurrency behavior for batch calls.
If our concurrency limit is 50 and we send a batch of 10,000 calls, will the remaining 9,950 calls be held in a queue and automatically dialed as each active call finishes? Or will they get rejected?
Please confirm that no calls will be dropped or rejected due to concurrency limits.