Call settings control how long an AloAi Voice Agent stays on a call and how it handles silence. Advanced call handling controls what an outbound agent does when a call does not reach a person right away: voicemail, a call-screening assistant, or an automated phone menu (IVR, interactive voice response).
Both are panels on the agent’s Configure tab: Call Settings and Advanced Call Handling. This article explains each setting and what the caller experiences.
To get there, open the Agents menu and click the agent’s name. The Call Settings and Advanced Call Handling panels are on the right side of the Configure tab.
Open the agent from the Agents list.
The Call Settings and Advanced Call Handling panels on the Configure tab.
Call settings
The Call Settings panel applies to inbound and outbound agents. It has these settings:
The Call Settings panel.
Voice AI Best Practices injects voice AI guidelines for natural speech, pronunciation, and safe conversations. It is turned on by default.
End Call on Silence ends a silent or idle call after the duration you set on the slider. It is Off by default, so the call does not end due to silence.
Max Call Duration sets the longest a call can last. It defaults to 15 minutes and can be changed per agent. When a call reaches the limit, the agent hangs up automatically.
On an outbound agent, if both End Call on Silence and AI starts speaking after silence are off, the editor shows a warning: if no one speaks, the agent waits in silence until the call times out, which can take several minutes.
To prevent this, turn on AI starts speaking after silence, or set End Call on Silence in Call Settings. For AI starts speaking after silence, see Writing instructions and setting conversation behavior.
Advanced call handling
The Advanced Call Handling panel appears on outbound agents only. Inbound agents do not show it, and voicemail detection, call screening, and IVR detection stay off for them.
The panel has three settings: Voicemail Detection, iOS Call Screen Handling, and IVR Detection. On a new outbound agent, Voicemail Detection is on, and the other two are off.
The Advanced Call Handling panel.
Voicemail
On each outbound call, the agent uses answering machine detection (AMD) to check whether a person or a voicemail system answered.
While the agent works out who answered, it stays quiet. It does not speak before the contact does, and it waits for more than a single opening word before deciding, because a person and a voicemail greeting often start the same way, for example “Hello.” A greeting that includes a voicemail phrase, such as “please leave a message after the tone,” is treated as voicemail.
What to expect on each type of call:
A person answers. The agent replies shortly after the person speaks, including when the person says only “Hello?”
The call reaches voicemail. The agent stays quiet through the greeting and then runs the voicemail action you chose.
The line plays a recorded notice first, such as a call-recording disclosure or a call-screening service. The agent stays quiet through the notice and replies once the person speaks. These calls take slightly longer than a call a person answers directly.
A tone on its own does not end a call that a person has answered. If the contact says a greeting such as “Hello?” two or more times, with pauses in between, before a beep-like tone, the agent speaks and waits for a reply. If the line stays silent, the agent runs the voicemail action.
When the agent detects voicemail, it waits for the greeting and the beep to finish before acting. If the voicemail system does not play a beep, the agent waits briefly and then continues, so it does not stay on the line.
With Voicemail Detection on, choose under Voicemail Response what the agent does when it reaches voicemail:
Voicemail Response with the Static Sentence message.
Hang up if reaching voicemail ends the call without leaving a message. This is the default.
Leave a message if reaching voicemail leaves one of these:
Static Sentence leaves a pre-written message, which can include dynamic variables such as {{FirstName}}. A default message is filled in, and you can edit it.
Prompt leaves a message generated from the instructions you provide, adapting it to previous conversations with the contact. Dynamic variables are supported.
To hear the voicemail message before going live, see Testing and reviewing your agent.
Call screening
Call screening helps outbound calls reach contacts whose mobile phones use a call-screening assistant, such as the iOS prompt that asks the caller to record their name so the phone can check whether the person is available.
When a screening assistant answers, Aloware recognizes the prompt and waits while the contact decides whether to take the call. Screening detection runs before voicemail detection. What happens next depends on the contact:
If the contact picks up, the agent is connected and the call continues normally.
If the contact does not pick up, the agent does not leave a voicemail by mistake.
Turn on iOS Call Screen Handling in the Advanced Call Handling panel to enable a predefined message for when call screening picks up. No per-call setup is needed.
IVR detection
IVR detection lets the agent recognize when an outbound call reaches an automated phone menu and navigate it until it reaches a person.
Turn on IVR Detection, then choose under IVR Response how the agent handles an IVR:
IVR Response with Custom guidance.
Hang up if reaching an IVR ends the call. This is the default.
Navigate the menu if reaching an IVR has the agent work through the menu until it reaches a person.
Custom guidance (optional) appears when you choose to navigate. Tell the agent who it needs to reach, so it picks the right option on each menu, for example which department to choose. It accepts up to 1,000 characters. Leave it blank and the agent works its way toward any live person.
During navigation, the agent tells apart automated menus, hold or queue announcements, and people:
Automated prompts are treated as menus, and the agent navigates them.
A short question from a person, such as “Hello?” or “Who is this?”, is treated as a person, and the agent switches to its greeting.
If an announcement says no one is available, the agent stops navigating instead of waiting indefinitely.
When the agent reaches a person, it delivers its greeting promptly and does not talk over the person if they speak first.