Language issues can affect HIV and AIDS research trial enrolment, but little is understood about variables in this process. Some evidence indicates barriers exist even when the trial enrollers are linguistically and culturally matched to the participants and when consent forms are translated into a local language. Counsellors from two sites of an antiretroviral (ARV) drug trial in South Africa were trained to enrol patients using a modified process of informed consent. The outcomes of the standard and modified processes were subsequently compared in regard to accuracy of the patients' recall, aspects of the content and process, and evaluations made by the patients and counsellors. Marked differences were noted in the effectiveness of the enrolment process after the counsellors' training, as reflected in the potential participants' scores for comprehension and their evaluation comments. A flexible approach to seeking enrolment which harnesses natural strategies of interaction has the potential to improve the process of trial enrolment and informed consent, and decrease participants' anxiety and uncertainty. The findings have implications for methods of trial enrolment and specifically draw attention to the potential role of language specialists in research ethics committees and in trial designs and monitoring.