Speech recognition software introduces audio and human interface into the equation, making this one of the programmes with the most moving parts. It’s in this light that we come across some of the most unique tech support queries, some that are almost impossible to address in a remote capacity. Going through a number of scenarios and rigorous testing of software and hardware can take some time. In order to shortcut the laborious process of investigating specific and singular issues, it’s usually best to nip things in the bud with a fresh installation.

While we would want to do our level best to ensure users aren’t hampered by tech issues, if we don’t have a file for an issue, it’s more prudent to start with a refresh rather than spending an inordinate amount of time trying to address what’s wrong. Even after a fresh install, which typically takes about 20 minutes, we may end up with the same error or issue. This is why we typically take a multi-prong approach to tech support, working by narrowing the problem through a process of elimination. Could it be the headset? Is the programme working successfully for the user on a second PC? Will a back up version of the user’s profile deliver different results? It can be a real hint hunt to get to the bottom of a tech issue.

In one case, after struggling to remedy a Bluetooth headset connectivity issue, we realised that the base unit wasn’t getting enough power. In order to neaten up loose cables, the user decided to buy a longer connection cable, which wasn’t adequately powered to run the system. Obviously, without this information it can be incredibly difficult to diagnose the problem unless the user brings it to our attention. However, gaining a degree of experience from resolving these issues, we do have a number of fixes and ideas of what could possibly disrupt a speech recognition software environment that was working previously with no issue.

So it comes down to trial and error sometimes, forcing our technicians to become like detectives in making their assessment. Running through a number of possibilities, interviewing the user in terms of what’s changed recently and getting a clearer picture of what could be wrong.

Trial and Error: Troubleshooting Dragon
Tagged on: