The purpose of this survey was to assess the workload of an emergency ambulance service, to describe the use of paramedic skills by those staff with full extended training, and to predict the impact upon the provision of pre-hospital care of deploying a paramedic on every emergency ambulance. Accordingly, a week-long survey was undertaken of all urgent and emergency calls received by an ambulance service covering a mixed urban and semi-rural area of 187 square miles with a population of 396,000. Of the total 682 emergency calls 351 (51.5%) originated from the '999' system: 291 of these patients were taken to hospital where 51% were thought to have minor conditions and 141 were admitted. General practitioners made 236 (34.6%) emergency calls: 234 patients were taken to hospital where 76.4% were thought to have potentially serious conditions or an acute risk to life and 217 were admitted. There was no difference in the type or severity of conditions attended by paramedic or non-paramedic crews. Time spent on-scene was significantly longer when paramedics were present (mean 11.0 min, 95% confidence interval 9.54-12.46 min v 8.31 min, 7.49-9.13 min) (p less than 0.01). Extended skills were used by paramedics in 42 (23.6%) of their patients, most of whom were medical cases. One patient was resuscitated from cardiac arrest. The presence of a paramedic on every emergency ambulance increases the time spent on-scene and offers advanced pre-hospital skills to patients who need them. Care should be taken to ensure that the benefits of time spent on-scene using such skills outweigh the disadvantage of delayed hospital admission.