A customer gives their name, provides an address or mentions a town. Your team member understands who or which location they mean, but the transcript may use a different spelling. These are precisely the details that matter when following up on the conversation.

A glossary gives AnswerPal extra context when transcribing audio recordings. It contains the correct spellings of proper names, addresses, towns and specialist terms that come up in your conversations, helping the transcription recognise those words.

Names and locations with the correct spelling

Consider an estate agency that frequently discusses properties in the same area. It can create a document containing contact names, street names and towns. An example list might look like this:

• Lotte Van den Broeck
• Karel Oomsstraat
• Sint-Katelijne-Waver
• ‘s-Gravenwezel

A list or table with the correct spellings is enough. There is no need to explain how each word is pronounced. Company names, product names and specialist terms can also be included in the same document.

Use the names and locations relevant to your organisation. This can be a more extensive list, for example covering the towns and streets in your service area.

Linking the document to transcription

You manage the glossary as a plain text document in AnswerPal. Use OnTranscribe to link it to recording transcription. AnswerPal then uses the linked glossary as additional recognition context.

1. Under Settings > Documents, create a text document with a clear name, such as Names and locations — our service area.
2. Go to Settings > Events. On the Hooks (Events) page, select Documents next to OnTranscribe.
3. Select your glossary and use Channels to choose where it applies: a single channel, a channel type or All Channels for the whole organisation.
4. Click Add to add the link. Then check a new recording to see whether the glossary fits your conversations.

Would you also like to use the glossary for an audio upload without an associated channel? Link it across the organisation using All Channels. Only organisation-wide documents are used for this type of upload.

For call recordings, audio attachments and uploads

The glossary can help when transcribing a full phone recording after the call, audio attachments in email and WhatsApp, and files uploaded through audio transcription. Short audio attachments can use it too.

The glossary is used when the recording is transcribed. Changes to the list are picked up by subsequent transcription jobs.

Check the results with your own recordings

Start with a few representative recordings and compare the transcript with the audio. Pay particular attention to names, addresses and towns that matter for follow-up. This helps you identify terms to add and check whether the spellings fit the way your organisation works.

There is a technical limit to the amount of text the transcription model can use. If this is exceeded, a Transcription notice appears and the recording is transcribed without the glossary. You can then adjust the list for future recordings.

Start with the words that make a difference

Gather the names, addresses, towns and specialist terms that are regularly misspelled in your transcripts. Keep the document up to date as new contacts, locations or products are added.

Would you like to set this up for your organisation? Contact AnswerPal. We can help you link the right document and plan a practical test using your recordings.