How throttling works

Feature deprecation: Genesys will no longer support ACD Web Chat v2, which is available to customers via all its corresponding Chat Widget versions. This is further to the deprecation of ACD Web Chat v1, which was announced earlier. As a result, Predictive Engagement will also end support for these legacy web chat versions. For more information, see Deprecation: Removal of ACD Web Chat (version 2). All existing customers are encouraged to migrate to Web Messaging and Messenger.

Note: This article applies only to customers using web chat. If you are a Genesys Cloud CX customer, we encourage you to use the new web messaging feature to replace web chat.

What throttling does

In Genesys Predictive Engagement, throttling is the process of limiting the number of web chat offers to protect your agent queues from being overwhelmed with an influx of proactive chat offers. Throttling doesn’t attempt to optimize for a specific metric or KPI. Rather, it reacts to signs of queue saturation and attempts to reduce the rate of incoming chats to mitigate the queue saturation.

How throttling works

Genesys Predictive Engagement monitors the states of chat engagements to track the service level observed on a queue. Genesys Cloud CX calculates the service level according to the standard Genesys Cloud CX formula. Genesys Cloud CX uses the observed service level values to predict the service level on the queue for the next minute. If the service level is predicted to fall below the configured service level for the queue, Genesys Predictive Engagement throttles the number of chat offers to improve the service level. Once the service level is predicted to be at or above the configured service level, Genesys Predictive Engagement reduces throttling and offers more actions.
Note:
  • Service level throttling only applies for web chat action maps when Route when agents are available is enabled on the action map.
  • The service level configured for the queue is not a target that throttling aims to achieve; instead, it’s the minimum under which throttling starts. Observed service levels on a queue may be above the configured service level.
  • Short abandons don’t count toward service level calculations for throttling purposes.

    Service level formula

    The service level formula is:

    (nengaged – noverSla) / (nengaged + nabandoned)

    nengaged = number of engaged actions

    noverSla = number of actions that take longer than the service level target value to go from started to engaged

    nabandoned = number of abandoned actions

    The service level target defaults to 20 seconds, but you can modify it in the queue configuration in Genesys Cloud CX. For Genesys Multicloud CX, use a public API to modify the service level target. For more information, see Update a single action target.

    Service level threshold

    The service level threshold defaults to 80 percent, but you can modify it in the queue configuration in Genesys Cloud CX. For Genesys Multicloud CX, use a public API to modify the service level threshold. For more information, see Update a single action target.