Select a region

region-selection
contact-us

Search Results

  • There are no results for ""

DingGo

about-banner

Automating Car Repair Decisioning through AI Damage Assessment and Cost Estimation

DingGo is a leading automotive crash management service provider operating across Australia and New Zealand, focusing on helping fleets minimise vehicle repair cost and time by streamlining all aspects of crash management across repairers, assessors, insurers and third party demands and recoveries.

With growing demand for instant and accurate crash assessments and the AI market rapidly maturing, DingGo set out to build a Crash Intelligence platform that would enable its fleet and retail customers to upload photos of damaged panels and – within seconds – receive a damage assessment report and a repair cost estimate, allowing them to decide how best to proceed with the repair.

To achieve this, DingGo partnered with Material to create a cutting-edge, modular, AI platform for automating car damage detection, severity scoring, panel classification and cost estimation.
  • Shortlisted and evaluated a set of computer vision AI models and architectures (including DingGo’s early home-grown proof of concept (POC) AI model) to identify the most effective solution for real-world damage detection, severity scoring and panel classification
  • Collaborated with DingGo stakeholders to define and prioritise features and improvements based on model evaluation findings
  • Developed image augmentation and enhancement functionality to convert pre-labelled damage images into AI-ready training data; the same dataset was also used to refine and retrain the selected model
  • Integrated Large Language Models (LLMs) to bootstrap the computer vision AI model with contextual accuracy, improving panel classification and damage severity categorisation
  • Developed a data-driven algorithm that factors in the client’s repair cost logic and industry standards to translate AI-detected damage into structured repair cost estimates
  • Implemented secure, role-based access controls to meet the needs of diverse user groups including clients, insurers, workshops and DingGo’s internal teams
  • Developed model observability monitoring within DingGo’s Snowflake Data Hub to let the client monitor live model performance and identify areas for model retraining
  • Deployed the complete AI solution as a scalable, API-first service on AWS and trained internal teams to operate and manage the solution independently
  • Developed an automated model retrain pipeline to enable DingGo to retrain the computer vision models with further labelled training data, and compare model accuracy

DingGo’s AI Crash Intelligence platform is currently being integrated within DingGo’s core operational workflow to transform damage inspection and repair cost estimation workflows.

The platform is showing strong early results, with damaged detection and classification exceeding 75% accuracy, and repair cost estimation exceeding 85% accuracy. It also gives DingGo the ability to throttle between full automation or have a level of human-in-the-loop review for accurate estimation.

Leveraging its modular API-first architecture, DingGo intends to monetise the Crash Intelligence platform to enable the brand’s ecosystem partners (assessors, repairers, insurers, dealers and third-party vehicle data platforms) to integrate instant damage assessment and repair cost estimation into their services. By unlocking accuracy, efficiency and scalability through AI, Material has empowered DingGo with a clear edge in Australia’s auto crash management market.

Have Questions? Get In Touch

CONTACT

More Client Stories

Elevating Digital Engagement for a Personal Care Brand Across ANZ With Drupal and Acquia

Transforming Vision Care Customer Journeys through Augmented Reality

Get in touch with us.

Let’s talk