Multi-access keys, such as those built using Lucid, have in their database character information about the taxa that are to be identified. When the user of a multi-access or matrix key chooses a character state of a selected character in the key, the taxa that have that character state are retained; taxa that have a different character state for the selected character are discarded. When another character is chosen, the same process is repeated with the remaining taxa. Since computers easily handle repetitive tasks at high speed, they are ideal for multi-access keys, which depend on a complete check of all the taxa in the key's database each time a question is answered.
Since multi-access keys allow you to start with any character you choose and to proceed in any order you choose, these keys do not suffer from the unanswerable couplet problem encountered in dichotomous keys. Further, in using a multi-access key you can avoid characters that are difficult for you to distinguish or are not appropriate for your specimen, another advantage of multi-access keys over dichotomous keys.
Ultimately, when using a multi-access key, all taxa except the one that matches your specimen may be eliminated. However, if you cannot eliminate all the taxa, you will be left with a small group of taxa that can then be compared more closely.