The ultimate aim of intervention for children with language impairment is an improvement in their functional language skills. Baseline and outcome measurement of this is often problematic however and practitioners commonly resort to using formal assessments that may not adequately reflect the child's competence. Language sampling, transcription and analysis provide a more thorough and realistic picture of a child's abilities but are time consuming and usually considered unfeasible in the typical clinic setting. This article reports on a pilot study in which a speech and language therapy assistant (SLTA) carried out language transcription using language analysis software. Following a brief initial training period, the SLTA carried out transcriptions on 17 language samples at two time periods. Reliability between a speech and language therapist and the assistant's transcriptions were calculated and were found to be acceptable for the majority of measures taken. The use of software by assistants is proposed as a viable alternative for outcome measurement of naturalistic language skills.