Speech data · Jordan

Jordanian Arabic call-center speech data

Customer-service conversations spoken the way Jordanians speak: Amman’s urban dialect, rural and Bedouin speech, and the Arabic–English mix of everyday calls. Every recording is transcribed word for word and checked by several native Jordanian reviewers.

Request samples Talk to us

Specification

Dialects: Ammani (urban), rural (Fallahi) and Bedouin Jordanian Arabic, with natural Arabic–English code-switching
Conversations: Customer and agent, two speakers per call: banking, telecom, insurance, e-commerce and government services
Audio: 8 kHz telephone or 16 kHz wideband, 16-bit WAV, mono or one channel per speaker
Speakers: Native Jordanian speakers, balanced across gender and age groups
Annotation: Verbatim transcript, timestamps, speaker turns, gender and dialect tags; intent, sentiment and entities on request
Quality check: Every transcript is cross-checked by several native Jordanian reviewers

Example of the data

Agent00:01
مسا الخير، معك رنا من خدمة العملاء، كيف بقدر أساعدك؟
Good evening, this is Rana from customer service. How can I help you?
Customer00:04
مسا النور، والله صرلي يومين بحاول أفعّل الباقة وما زبطت معي.
Good evening. I’ve been trying to activate the bundle for two days and it won’t work.
Agent00:09
ولا يهمك، بس أعطيني رقم الخط لو سمحت.
No problem, just give me the line number, please.
Customer00:12
أكيد، بس قبل هيك بدي أعرف ليش انخصم مني مبلغ زيادة.
Sure, but first I want to know why I was charged extra.

What it is used for

Speech recognition for Jordanian calls
Train and evaluate ASR that hears Jordanian dialect, not only Modern Standard Arabic.
Voice bots and IVR
Let callers speak naturally instead of pressing buttons or switching to formal Arabic.
Call analytics
Detect intents, complaints and sentiment across every call, in dialect.
Agent quality monitoring
Score calls automatically against your service standards.

Questions

Which Jordanian dialects does it cover?

Amman’s urban dialect, rural (Fallahi) speech and Bedouin speech, plus the Arabic–English code-switching common in Jordanian calls. Tell us the mix your product needs.

What annotation comes with the audio?

Verbatim transcripts with timestamps, speaker turns, speaker gender and dialect tags. Intents, sentiment, named entities or your own label set can be added.

Can I hear samples first?

Yes. Request samples with the form on this page and we will send them with the full specification.

Can it be used in a commercial product?

Yes. The data is prepared for training and evaluating commercial AI systems, under a licence agreed with you.