Build a Buyer Complaint Audio Transcription and ONDC Issue Drafter
Convert buyer complaint voice recordings in regional Indian languages into formal ONDC issue payloads. Map spoken grievances into standard network codes for seller dispute routing.
What you are building
The core problem, expected build, and operating context for this challenge.
Convert buyer complaint voice recordings in regional Indian languages into formal ONDC issue payloads. Map spoken grievances into standard network codes for seller dispute routing.
E-commerce customer service triage agent
Process customer voice notes through simulated Bhashini ASR and NMT pipelines.
Extract asserted defect facts, order reference IDs, and customer resolution desires.
Map vernacular expressions of dissatisfaction to standardized ONDC grievance issue subcategories.
Produce formatted Beckn issue transaction payloads for customer service agent sign-off.
Grievance listening console with waveform playback, bilingual text view, and Beckn issue draft panel.
Generate synthetic voice transcripts of buyer issues paired with ONDC issue logging taxonomy tables. Starter examples below are synthetic. They are not downloaded records.
ONDC protocol specifications Adopts ONDC protocol draft specifications for buyer-seller issue logging and dispute arbitration.
semantic-kernel, pydantic, fastapi. Enables native integration of multi-step linguistic translation and structured template generation. Claude API or high-capability open-weights model for dialect idiom interpretation.
Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.
Transcribe regional audio into standard English issue summaries accurately.
Assign the correct ONDC issue subcategory based on stated complaint criteria.
Extract order identifiers and item details into standardized message structures.
Prevent automatic network dispatch until customer care staff confirms accuracy.
```json [ { "input": { "complaint_id": "CMP-302", "voice_text": "Mera dabba khula hua mila aur andar ek botal gayab thi order number ONDC-774", "source_language": "hi" }, "expected_output": { "extracted_order_id": "ONDC-774", "issue_category": "ORDER", "issue_sub_category": "ORD02_ITEM_MISSING", "summary_en": "Package arrived unsealed with one bottle missing from the order.", "beckn_action": "issue" } }, { "input": { "complaint_id": "CMP-911", "voice_text": "Delivery wala banda badtameezi se baat kar raha tha", "source_language": "hi" }, "expected_output": { "extracted_order_id": null, "issue_category": "FULFILLMENT", "issue_sub_category": "FLM04_AGENT_BEHAVIOR", "summary_en": "Delivery agent behaved unprofessionally.", "beckn_action": "issue", "missing_order_reference": true } } ] ```
Does not penalize delivery agents or escalate disputes automatically without human agent oversight. Use mock external writes. Require human review before any real deployment.
Participation guide
Does not penalize delivery agents or escalate disputes automatically without human agent oversight. Use mock external writes and human review.
How work is evaluated
Transcribe regional audio into standard English issue summaries accurately. Assign the correct ONDC issue subcategory based on stated complaint criteria. Extract order identifiers and item details into standardized message structures. Prevent automatic network dispatch until customer care staff confirms accuracy.
Runnable source, pinned dependencies, setup commands, working interface, example outputs, recorded demo, tests, and action logs.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Complete workflow
Process customer voice notes through simulated Bhashini ASR and NMT pipelines. Extract asserted defect facts, order reference IDs, and customer resolution desires. Map vernacular expressions of dissatisfaction to standardized ONDC grievance issue subcategories. Produce formatted Beckn issue transaction payloads for customer service agent sign-off.
Missing or incorrect.
Partly complete, with documented gaps.
Complete, with reproducible evidence.
Correctness and evidence
Transcribe regional audio into standard English issue summaries accurately. Assign the correct ONDC issue subcategory based on stated complaint criteria. Extract order identifiers and item details into standardized message structures. Prevent automatic network dispatch until customer care staff confirms accuracy.
Missing or incorrect.
Partly complete, with documented gaps.
Complete, with reproducible evidence.
Review interface
Grievance listening console with waveform playback, bilingual text view, and Beckn issue draft panel.
Missing or incorrect.
Partly complete, with documented gaps.
Complete, with reproducible evidence.
Safe execution
Does not penalize delivery agents or escalate disputes automatically without human agent oversight.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Process customer voice notes through simulated Bhashini ASR and NMT pipelines.
Extract asserted defect facts, order reference IDs, and customer resolution desires.
Map vernacular expressions of dissatisfaction to standardized ONDC grievance issue subcategories.
Produce formatted Beckn issue transaction payloads for customer service agent sign-off.
Reference links and supporting material
Generate synthetic voice transcripts of buyer issues paired with ONDC issue logging taxonomy tables. Supplied starter examples are synthetic, not source extracts.
Challenge setup
Configuration summary for this challenge.
Challenge input
Input details are not configured.
Semantic Kernel
Agent Systems · Agent Frameworks
Evaluated output
No output checks are configured.
- Process customer voice notes through simulated Bhashini ASR and NMT pipelines. Extract asserted defect fact...
- Transcribe regional audio into standard English issue summaries accurately. Assign the correct ONDC issue s...
- Grievance listening console with waveform playback, bilingual text view, and Beckn issue draft panel.
- 2 public reference cases
- 1 configured tool
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Semantic Kernel · Optional
Evaluation contract
- Complete workflow · Weight 40
- Correctness and evidence · Weight 30
- Review interface · Weight 20
- Safe execution · Weight 10
Reference cases
- 2 public reference cases
Workflow tasks
No workflow tasks are configured.
Recipe freeze marker
No recipe freeze time is recorded.