Returning to FIG. 1, the communication service 108 may request a trust score from the scoring agent 110 and evaluate the trust score returned by the scoring agent 110 to determine whether to allow a connection request associated with a customer account. A trust score may comprise any type of indicator or set of indicators used to represent a risk of allowing a caller to connect to a destination endpoint. For example, a trust score may be a binary value, classification value (e.g., Yes or No or weighted values), numeric value, or a set of values that can be evaluated by the communication service 108 to determine whether to allow or deny a connection request. The communication service 108 may evaluate the value of the trust score and either allow or deny a connection request based on the evaluation. As an example, the scoring agent 110 may return a trust score classification indicating that a connection request should be allowed, and the communication service 108 may allow the connection request to be executed. As another example, the scoring agent 110 may return a numeric value, and the communication service 108 may evaluate the trust score and determine that the numeric value meets a threshold that allows the connection request to be executed. In yet another example, the scoring agent 110 may return separate trust score values for a number of trust categories (e.g., destination risk, customer status, call history, and other categories) to the communication service 108, and the communication service 108 may evaluate the trust score values and determine whether the trust score values meet a threshold that allows the connection request to be executed. For example, in the event that a majority of categories have assigned trust score values that allow the connection request, the communication service 108 may allow the connection request. In some examples, the separate trust score values returned by the scoring agent 110 may be weighted according to a weight assigned to a trust category.