If your business has a significant outbound calling requirement, utilizing AI-powered outbound call tasks can significantly improve the efficiency and productivity of your agents.
Key benefits of using AI outbound call tasks:
- Automated dialing at an optimal rate based on algorithmic capabilities.
- Filtering of invalid numbers, voicemail, and no-answer calls to focus on high-value connections.
- Identification of live human answers and seamless transfer to agents to complete the conversation.
1. About AI outbound tasks
The AI outbound call process is as follows:
(1) After uploading the numbers to the task, the AI automatically dials the numbers.
(3) The AI filters out invalid numbers and automatically adds AI intent labels:
- Filter out numbers with format errors or invalid numbers
- Filter out unreachable invalid numbers
- Filter out voicemail (can attempt multiple redials for voicemail numbers)
- Filter out unanswered calls (can attempt multiple redials for unanswered numbers)
(4) When the AI detects a live answer, it immediately transfers the call to an available human agent:
- If there is an available agent (online and idle), it immediately transfers the call to that agent
- If there are no available agents (offline or busy), it puts the caller in a queue and plays a hold message while waiting for an agent
(5) The agent answers the call, completes the conversation with the customer, and records the call details.
(6) Before the agent answers: if the queue time expires or the customer hangs up, it results in a lost call.
(7) The system adjusts the dialing rate based on lost call rate, agent utilization, connect rate, and average call duration (only for predictive dialing)
The call states in the AI outbound ：call status are determined based on the feedback from the telecom carriers and the call routing center.
| Pending | |
| Scheduling | |
| Dialing | |
| Ringing | |
| Talking | |
| Talked | |
| Invalid Number | |
| Failed to Connect | |
| User Occupied | |
| User Busy And Rejects | |
| User Rejected | |
| Scheduling Error | |
| TTS Processing | |
| TTS Failed | |
| Unknown | |
| Blacklisted Number | |
| Stopped | |
| Not sufficient funds | |
| Invalid Number Direction | |
| Invalid Number | |
| Ring missed | |
| No response to call | |
| Morning Media - Voice Mail | |
| Ring too long | |
| No line available |
AI outbound Tags :Tags are based on the call status and AI recognition capabilities.
| Pre dialling - invalid number |
| Pre dialling - customer suspended |
| Pre dialling - no availabel line |
| Dialing - dailout fail |
| Call Engagement - unknown cause |
| Call Engagement - no response |
| Call Engagement - user is busy and rejects |
| Call Engagement - user refuses to answer |
| Call Engagement - incorrect number |
| Call Engagement - the user is busy |
| Call Engagement - ring but no answer |
| Call Engagement - ringing timeout |
| Call Engagement - voicemail before answer |
| Connection - user hangup |
| Connection - system error |
| Connection - mute |
| Connection - hang up immediately |
| Connection - voice mail |
| Connection - transfer fail-user hangup-waiting |
| Connection - transfer fail-user hangup-ringing |
| Connection - transfer fail-user waiting timeout |
| Connection - transfer success |
2. Create an AI Outbound Task
Operation：Call center-AI outbound task-new task
Parameters:：
- Task Name: Enter the task name
- Task Type:
- Proportional Outbound call : After setting the ratio, the system will automatically dial calls based on the fixed ratio according to the number of available agents.
- Predictive Outbound call : After setting the ratio, the system will initially make automatic dial-out calls based on the fixed ratio according to the number of available agents. After dialing a certain number of numbers, the system will automatically adjust the call rate based on the statistical results, the agent utilization rate, the agent call duration, and other data to balance the task speed, agent utilization rate, and call loss rate.
- Agent Call Rate:
- The fixed ratio for Proportional Outbound
- The initial ratio for Predictive Outbound
- Ringing time: the max ring time
- Country/Region: The country/region that the outbound task is targeting
- Dialing Strategy: Specify the replay strategy for unanswered calls and voicemail, with a maximum of 4 replays allowed.
- Agent Group: When the outbound call is answered, the call will be transferred to this agent group.
- Call Routing: If you have a self-built routing, you can select it, otherwise, choose the default routing.
- Import Numbers: You can batch upload the numbers to be dialed based on the number information template.
3.Task Control
(1) You can start, pause, and delete the task at any time.
(2) During the task execution, you can switch the task type and adjust the dialing multiplier.
- If the number dialing rate is too fast:Reduce the dialing rate
- If the dialing rate is too slow: Increase the dialing rate
- If there are a large number of unexpected call losses: Switch the task type to Proportional Outbound and reduce the dialing rate.
4.Task list
| Key metrics | Explain |
| Completed|Numbers|Attempts | Completed: The quantity of numbers that have been connected or completed all attempts Numbes: The quantity of numbers uploaded in the task Attempts: The quantity of dialing attempts |
| Transfer%｜ Answer(numbers) % | Answer(attempts)% | Transfer%: number of calls transferred to the queue wait for agents ÷ number of calls answered by users. Answer(numbers) %: number of calls answered by users ÷ numbers Answer(attempts)%: number of calls answered by users÷ attempts |
| Loss% | Loss | Agent talk time + after-call work time / (Agent talk time + aftert call work time + idle time) -- All agent in group |
5. Details-Monitoring
Clicking on the task details allows you to view the task monitoring, including:
- Completed | Numbers | Attempts
- Progress: Completed / Numbers
- User answered: The number of users connected (including voicemail)
- Transfer%|Answer(number)%|Answer(Attempts)%|
- Loss% | Loss
- Scheduling: Displays the system's scheduling status in the past 10 minutes. Busy if agent utilization is too high, idle if too low, otherwise normal.
- Utilization: The agent utilization rate in the past 10 minutes
- Task Monitoring: Historical trend charts for the above metrics
- Agent call duration Distribution: Distribution and proportion of talk time
- Scheduling Monitoring: Detailed view of agent status and call scheduling in the past 24 hours
6. Details-All numbers
The report records are on a per-number basis, displaying the dialing results for all uploaded numbers. Each number will have one and only one final result. This report is suitable for focusing on the final communication outcome of each number.
| Key items | Explain |
| Completion status | Completion status: Whether the number has finished the diale and its current status, include: pending dialing, scheduling, dialing, ringing,in conversation, and completed |
| Answer status | Result of whether the number was answered by the user |
| AI tag | Dialing label automatically identified by AI |
| Wrap-up tag | Label assigned by the agent after the conversation |
| Attempt strategy | Attempt strategy for the number |
| Rounds | Dialing rounds experienced by a number |
7. Details-All attempts
The report records the result of each individual dialing attempt, suitable for observing the data on a per-dialing basis. This is in contrast to the previous report, which focuses on the final communication outcome for each number.
| Key items | Explain |
| Status | Call status of the current dialing |
| Recording | Recording starts from the user is connected (including voicemail) |
| Agent Group | The agent group to which the call is assigned |
| Agent | The agent to which the call is assigned |
| AI tag | AI-recognized call tags |
| Wrap-up tag | The tag that the agent artificially records for the call after communicati |
| Attempt strategy | Attempt strategy for the attempt |
| Rounds | The round of this call |
8. Historical Report for AI Task
The data records under the AI Outbound Calling Task menu are task-level data statistics. The historical data statistics for all AI tasks can be viewed in the Reports - Contact Center Reports - AI Calling Report.