Reviews have demonstrated large effect sizes when using computerised cognitive behaviour therapy (CBT) protocols for treating anxiety, depression and health related concerns. However, the amount of therapist contact per user seems to be the most significant prognostic indicator. Thus, in some ways current online interventions can be viewed primarily as an extension of one-on-one therapy. The present article provides guidelines for targeting this limitation of online psychological interventions. The goal is to mimic the therapeutic relationship using a computer, without having any therapist involved. Consequently, thousands of users would be able to receive treatment simultaneously, reaching a wider audience, which was the initial goal of the online model. The development of a treatment program using file audit data is suggested as an alternative to having an individual therapist for each user. This is done by allowing the ‘computer psychologist’ to tailor individualised treatments for each user based on their psychological profile. The user is provided with individualised corrective feedback based on a set of prewritten responses to common faulty thoughts. A new paradigm is proposed for online treatment delivery.

This paper introduces a novel approach to internet treatment for social anxiety. The goal of this treatment was to address key limitations of current standalone treatments (Helgadottir, Menzies, Onslow, Packman, & O'Brian, 2009). The ‘computer psychologist’ designed for this study used fully automated, prewritten individualised sample answers in order to simulate a human–human interaction through a human–computer interface. Two males who sought treatment for stuttering and met the diagnosis for social phobia according to the DSM-IV and ICD-10 criteria were selected for this study. After receiving the treatment, both users no longer met criteria for social phobia. Also, significant improvements were observed on other psychometric tests, including measures of unhelpful cognitions, behavioural avoidance, quality of life, and low mood. The quality of the interaction appeared to be similar to face-to-face therapy, indicating that the ‘computer psychologist’ established an effective therapeutic relationship, and the automated techniques used were sufficiently engaging to prompt users to log on regularly and complete the treatment program.