Advanced AI Transcription and Summaries

AI transcription automatically turns the audio of a phone call into written text. On top of that transcript our AI also creates a summary: a short text covering why the call happened, what was decided and which action points follow. So you get both the literal conversation and its essence in a few sentences.

Under normal conditions accuracy is above 95%. Line quality, background noise, people talking over each other and heavy dialect can affect it. For technical terms, brand names and product codes you can have a custom vocabulary added so those terms are recognised correctly.

Dutch and Flemish, French, English and German are fully supported, alongside dozens of other languages. The language is detected automatically, which matters when your customer base is multilingual. The summary can be produced in the language of the call or in your organisation's language.

Yes. Speaker recognition splits the conversation per speaker, so you clearly see what the customer said and how your employee replied. Each passage also gets a timestamp, letting you jump straight to the right moment in the recording.

You choose: by email to one or more recipients, straight into your CRM, ERP or helpdesk, through an API or webhook into your own software, or simply in the Fonzer platform alongside your call history. Combinations are possible too, for example an email to the employee plus a note on the customer record.

Yes, provided your software allows it. We write the summary as a note or call log on the right contact, ticket or case file, using your package's API. If that package offers no suitable API or no field to store text in, we deliver by email or webhook instead. We check what is possible with you beforehand.

Fonzer already integrates with a range of common CRM, ERP, helpdesk and case management packages, and that list keeps growing. If yours is not on it, a connection through API or webhook is usually still possible. Have a look at our integrations page or ask your Fonzer partner.

No, you set the rules. Transcription can be enabled for specific users, queues, numbers or call types, for example only inbound calls to your customer service. Call recording has to be active for it, since without a recording there is no audio to process.

Processing starts the moment the call ends. For an average conversation the summary is in your mailbox or in the connected package within a few minutes. Very long calls take a bit longer, but in practice we are always talking minutes, not hours.

The audio and the text are processed on our own European servers and are not used to train external AI models. You sign a data processing agreement with Fonzer, access is limited to the users you designate, and traffic is encrypted. That keeps the processing within the GDPR framework.

By default we keep call recordings, transcripts and summaries between 30 and 365 days. On request that retention period can be adjusted to your policy or to sector obligations. Once the period ends, the data is deleted or anonymised.

Yes. Everyone taking part in the call must know it is being recorded and processed. You can do that with an automatic announcement at the start of the call, a mention in your privacy statement and in your waiting message. Fonzer provides the announcements, you decide the wording.

Transcription and summaries are an option on top of your Fonzer telephony. The price depends on the number of users, your call volume and the delivery you want. Your Fonzer partner will put together a proposal that matches your situation.

In three steps: activate call recording for the users or queues you want, choose where the summary should go and, if needed, set up the connection with your software. Most customers are up and running within a single appointment. Request a demo and we will walk through it together.