Speech to Text:
Recordkeeping is an integral part of the legal industry. Attorneys must document meetings for ongoing client cases, to fulfill increasing requirements for billing, as a deterrent against claims of fraud and abuse, and for malpractice protection. With the typical legal encounter often lasting less than 30 minutes, only a fraction of this time on average can be dictated for documentation.
Dictation, whether transcribed or speech recognized, is the natural way to record notes and is one of the most efficient ways to generate voluminous, granular, detailed, flexible and legible notes in real time or batch mode. Typically, even the most complicated meeting documentation can be accomplished during two or three minutes of dictation. Most attorneys dictate client meeting summaries as well as consultation letters.
Many versions of speech recognition software exist, however can lend confusion and disappointment to those who acquire home or retail versions, and try to make them work in an exacting environment requiring professional features. The same comments apply to the microphone that is usually included with the software. These must be professional units that have noise canceling mics - one facing the user and the other facing the opposite direction to sense ambient background noise.
We have spoken with many attorneys who state that they have this brand or the other, and when asked if they are using it they respond that: 1) "It wasn't accurate enough for my work", or 2) "It took too long to use or train". These are the typical reactions to the past experiences of many users. In our opinion, the retail market has been flooded with products that some users are trying to use for very professional applications. We feel this has resulted in adverse reactions and therefore you will hear so called experts say that "speech recognition is not quite ready for prime time". As you read further, you will see that this is far from the truth.
Performance Accuracy Considerations:
The following is a discussion of the factors that affect speech recognition performance and accuracy when used with speech recognition software. Accuracy can be defined as the inverse percentage of word and punctuation misrecognitions. For example, if a user's text has 4 misrecognitions out of a total of 100 words and punctuation, then the accuracy rate would be 96%. Some of the factors that impact successful performance and accuracy are:
- User pride and commitment in achieving high dictation accuracy.
- User enunciation skill and minimal regional, foreign or ethnic accent.
- User acceptance of the coaching/training direction by our trainers. All training is one-on-one.
- User experience with MS Windows and general computer use.
- A user’s vocabulary uniqueness.
- Training of the users on the techniques required to improve accuracy through continued use.
- Training to permit users to add new vocabulary terms, pronunciations and shortcuts.
- User dictation skill as well as preparation/organization before beginning to dictate
We are prepared to assure any new client that if the above procedures are followed a 96% accuracy rate can be achieved. We can demonstrate dictation of your documents (sight unseen) at an accuracy rate of 98% to 99%.
Naturally, the success of a new technology comes down to each individual's desire and motivation. Like any new methodology, past habits can control results and outcomes. We have had positive feedback from many of our test subjects who share the realized financial savings of using new technology and are embracing its use. Also, when an attorney's dictation is completed within minutes, without any misspellings or proofreading by others, there can be many tangible benefits to all concerned.
