Dear Community,
We have to solve the following requirement:
-
There are two different call center service providers for a service at the customer (one call flow for one service hotline)
1a. For example, assuming 100 percent for the total emerging volume, if call center service provider 1 gets 70 percent, then call center service provider 2 gets 30 percent -
The distribution of the calls should be configurable so that a supervisor can control this via UI without having to administrate a call flow.
-
The expected total call volume in absolute numbers for a service hotline is known
-
Each service provider is guaranteed an agreed absolute number of interactions in advance for their capacity planning.
Our approach so far looks like this:
-
There is a data table with one row per service
1a. Contents are: name service, percentage of calls call center service provider 1
1ai. The percentage of call center service provider 2 results automatically (100 - percentage of call center service provider1)
1b. Content is ad hoc customizable by a supervisor with appropriate permissions via UI. -
In the call flow (one per service) we generate a random number between 0 and 1 using the default random function,
2a. 0.15 would be 15 percent
2b. Based on a sufficiently large number of interactions, evenly distributed random numbers between 0 and 1 are generated -
When dividing the interactions between the two service providers, the generated random number is checked and compared to the data table
3a. Example 1 for a call processing
3ai. Datatable for service A envisages a load of 70 percent of the incoming calls for call center service provider 1, and a target value of 30 percent automatically results for call center service provider 2
3aii. Interaction reaches the call flow via a call route
3aiii. Caller selects the appropriate destination branch in the IVR menu
3aIV. Random number generated: 0.85
3av. Random number is greater than value for call center service provider1 (0.7)
3avi. Interaction is queued for call center service provider2 to the ACD for distribution.
3b. Example 2 for a process
3bi. Datatable for service B envisages a load of 40 percent of the incoming calls for call center service provider 1, and a target value of 60 percent automatically results for call center service provider 2
3bii. Interaction reaches the call flow via a call route
3biii. Caller selects the appropriate destination branch in the IVR menu
3biv. Random number is generated: 0.25
3bv. Random number is less than value for call center service provider1 (0.4)
3bvi. Interaction is queued for call center service provider1 to the ACD for distribution.
Over the period of 3 hours, we already achieve the target values of the percentage distribution with the existing call volume very well (approx. 2 percent deviation). Now we need a way to control the distribution of calls between call center service provider 1 and call center service provider 2 in a shorter period of time, ideally every 30 minutes. Due to the insufficient number of calls at such a clock rate, our approach produces rather blurry results (deviation of up to 7 percent). This deviation has the negative effect that the volume of the guaranteed calls (derived from the forecast of the known estimated call volume for a service) cannot be met for the Call center service providers in some cases and accordingly agents are sitting around without a workload.
Our idea of avoiding idle agents is to check for activity before distribution to the relevant target queues. However, this does not yet ensure that we will achieve the target quota in absolute figures.Do you have an idea how the interactions could be distributed for such a short period of time, so that no agents sit around without a workload and the committed calls can be assigned per call center service provider? Do you perhaps have a completely different approach that you would use to cover this use case?
Kind regards
Jan