The JET device is equipped with two Neutral Beam (NB) heating systems, each with eight Positive Ion Neutral Injectors (PINIs) arranged in four pairs, or "quadrants", each with its own ion deflection magnet [1]. Two configurations of PINI are used, for acceleration voltages up to 85kV and 160kV respectively, and there is presently one entire NB system in each configuration. Both injection systems were modified for compatibility with tritium operation in preparation for JET's first full DT experimental campaign ("DTE1"). This involved a number of mostly minor design changes to the beamlines (such as use of all-metal seals, provision of pumped interspaces, etc.) [2]. The most extensive modification was the implementation of a Tritium-Deuterium Gas Introduction System (TDGIS), which is supplied from the central Active Gas Handling System (AGHS) [3], and delivers gas via a single ground-potential inlet on each PINI. The TDGIS, its incoming gas transfer lines and the AGHS share a common secondary containment envelope. The design, commissioning and initial operating experience of the TDGIS have been described in detail in [4]. Only the 160kV injector system was commissioned in tritium; the 160kV PINIs had been designed to operate using the full capability of the high-voltage power supplies (160kV/30A) in tritium, and were expected to deliver up to 12MW tritium beam power to the plasma. The other injection system can deliver up to 13MW at 80kV/55A in deuterium. This combination of tritium and deuterium beams is the most favourable in terms of total delivered power and also results in a reasonable balance in the relative D:T beam fuelling rates at full power. The flexibility to allow individual beamline quadrants to operate in different gases (deuterium or tritium) simultaneously was included in the design of the TDGIS; the problem of interaction between adjacent beamline deflection magnets, when set for different beam masses, had to be addressed and is discussed in detail in an accompanying paper [5]. The option to commission the 85kV injector in tritium was retained only as a backup in case (for example) it was required to achieve satisfactory plasma D:T mixture control; it did not however prove necessary to exercise this option. In the present work, the experience obtained over the different phases of commissioning and operation is presented and discussed: specific tritium-related commissioning using deuterium; initial commissioning of tritium beams; routine tritium beam operation and beamline de-tritiation after DTE1. An intervention to repair the tritium beamline, which interrupted the programme described in this work, is discussed in a separate paper [6]