The project aims at producing multiservice Quality of Experience (QoE) estimator agents representing a human user. The estimators will be based on models of human perception and upon service discovery, trained progressively from the best-fitted current profile for that service. Each individual agent will communicate with other user agents through the network, exchanging profile information so that services and applications could adapt to both-end preferences and expectations on a peer-to-peer basis. Also, by participating in the agent’s network, the service provider can obtain QoE-relevant user information (by groups or individuals) so as to be able to adapt the resource provisioning from the network and thus optimize both the user experience and the cost of operation (e.g. in a cloud paradigm: the amount of computing resources, the bandwidth allocated to a particular service for a user or group of users, the physical resource location, etc.) By having a good estimation of user perception, the cloud can also better define the objectives of the service being provided, and translate them to the more technical objectives of the service components that need to be provisioned. Also, evolved Service Level Agreements (SLA) can be defined in terms of this QoE so that both the provider and the user have a better way of following up and measuring if the objectives are being met.