Evaluation and routing methods


PureCloud ACD uses a queue’s routing and evaluation methods to determine how to match interactions and agents. The evaluation method determines how PureCloud processes skill requests for interactions, while the routing method determines which of two routing types to use. A customer can request skills, for example, from the IVR. The evaluation method and the routing method for the queue determine the way that PureCloud uses the requested skills to match the interaction with an agent.

A queue administrator assigns an evaluation method and a routing method to each queue.


Agent profiles include ACD skill tags to indicate an agent’s areas of expertise and knowledge. A skill rating indicates the level of proficiency for each agent ACD skill. PureCloud uses agent skills and ratings for some ACD routing methods.

Note: When determining which agent to route the next interaction, PureCloud only looks back 7 days. If an agent has not been on queue in more than 7 days, PureCloud does not take into account that agent’s last handled time. For example, an agent who has not taken a call in 5 days will receive the interaction before an agent who hasn’t taken a call in 10 days.

The evaluation methods are:

Method Description Notes
All Skills Matching ACD only considers agents who have all of the requested skills. PureCloud looks for the first available agent with the requested skills.
Best Available Skills ACD only considers the first 100 available agents who have all of the requested skills and the highest average proficiency rating. PureCloud evaluates the first 100 agents to find the agent with the highest average proficiency rating. The average is calculated using the agent’s proficiency rating for each of the requested skills.
Disregard Skills, Next Agent ACD routes interactions based on agent availability instead of skills. PureCloud looks for the first available agent and ignores any skill requests.

In all three methods, the last-handled ACD interaction determines the order in which PureCloud evaluates the agent. PureCloud selects the agent that matches the evaluation method and has gone the longest since handling an ACD interaction. Status changes and on-queue behavior do not affect calculation and selection.


PureCloud ACD uses a queue’s routing and evaluation methods to determine how to match interactions and agents. The evaluation method determines how PureCloud processes skill requests for interactions, while the routing method determines which of two routing behaviors to use.

The routing methods are:

Method Description
Standard ACD Routes interactions to the next available agent. Considers skills as specified by the evaluation method.
Bullseye Routes interactions to a targeted sub-queue of agents with specific skills, and if no agents are available, it relaxes the requested skills, based on the queue configuration, to expand the pool of agents.



Together, the combination of evaluation and routing methods determines specific behavior.

Standard ACD Routing Bullseye Routing
All Skills Matching
  • Looks in the queue for the first available agent with all of the requested skills.
  • Evaluates agents in the order in which they became available.
  • If no agents with all skills are available, then the call waits until an agent with all skills becomes available.
  • Looks first for an agent with the exact skills in the targeted sub-queue.
  • Expands the selection pool of agents one ring at a time, either immediately or after some time delay, until it finds a match.
  • If no matching agents are available in the queue after the last ring is added, then the call waits until an agent with all of the skills becomes available.
Best Available Skills
  • For the first 100 available agents in the queue, it averages the agents’ skill ratings for the requested skills to find the best qualified available agent.
  • Evaluates agents in the order in which they became available.
  • If no agents with all skills are available, then the call waits until an agent with all skills becomes available.
  • Looks first for an agent with the exact skills in the targeted sub-queue.
  • Evaluates each available agent in the targeted sub-queue by averaging the agent’s skill ratings for the requested skills to find the best qualified, available agent.
  • Expands the selection pool of agents one ring at a time, either immediately or after some time delay, to increase the number of agents that are evaluated.
  • The evaluation stops when the first 100 agents of the selection pool have been checked and if a matching agent is found.
  • If no matching agents are available in the queue after the last ring is added, then the call waits until an agent with all of the skills becomes available.
Disregard Skills, Next Agent
  • Looks for the first available agent in the queue.
  • Evaluates agents in the order in which they became available.
  • If no agents are available, then the call waits until an agent becomes available.
  • Looks for an agent in the targeted sub-queue.
  • Expands the selection pool of agents one ring at a time, either immediately or after some time delay, until it finds an available agent in the queue.
  • If no agents are available in the selection pool, then the call waits until an agent becomes available.