In one example of the technology, a communication service may be configured to allow or deny a connection request to initiate a connection with a destination endpoint based on a trust score generated for a customer account. In one example, the communication service may receive a connection request associated with a customer account to initiate a connection (e.g., a voice, video, text, or data connection) with a destination endpoint. In response to receiving the connection request, the communication service may identify a customer account associated with the connection request and obtain a trust score (e.g., from a scoring agent) which indicates whether the connection request associated with the customer account is allowed. A trust score may be determined based in part on destination risk information, customer status information, customer call history information, and other information linked to the customer account, as described in greater detail later. In the case that the trust score indicates that the customer account is allowed to request the connection with the destination endpoint, the communication service may initiate the connection with the destination endpoint. In the case that the trust score indicates that the customer account is not allowed to request the connection with the destination endpoint, then the communication service may provide an indication or message sent to the caller that the customer account is not allowed to request a connection with the destination endpoint.