BMW speeds car development with a new app for defect ticket routing (PRO201)

Generative AI to the Rescue: Improving Ticketing Experiences at BMW

Overview

  • BMW has a long history of being an innovative company, and they are now exploring the use of generative AI to improve their ticketing process.
  • The goal is to address common issues with ticketing, such as inconsistent quality, unclear responsibility, hidden dependencies, and duplicate tickets.

Ticket Creation and Resolution Process at BMW

  • BMW has a well-defined software development process, with a funnel that starts with highly scalable and automatable unit and integration tests, and progresses to more complex testing in the cloud and in the vehicle.
  • Whenever a developer, tester, or other stakeholder identifies an issue, they create a ticket, leading to a large number of tickets that need to be processed.

Challenges with Ticketing

  • Tickets can have inconsistent quality, with gaps between the developer's and tester's understanding of the issue.
  • It's often unclear who is responsible for a particular issue, as there can be hidden dependencies between different components.
  • Duplicate tickets are a major problem, as they lead to wasted effort and frustration.

Leveraging Generative AI to Improve Ticketing

  • The team at the Generative AI Innovation Center worked with BMW to build a solution that leverages generative AI to address these challenges.
  • The solution consists of two main components:
    1. Ticket Enrichment: An LLM is used to summarize the ticket, eliminate acronyms, and translate the content to a common language.
    2. Duplicate Detection: A hybrid search approach, combining keyword-based and semantic vector-based search, is used to identify potential duplicate tickets. An LLM is then used to classify the tickets and provide an explanation for the decision.

Key Highlights

  • The enrichment process improves the quality and consistency of the ticket information, making it easier for the right team to be identified and routed.
  • The duplicate detection process significantly reduces the amount of duplicated work for developers, saving time and improving efficiency.
  • The addition of explanations from the LLM helps the human decision-maker understand the reasoning behind the duplicate detection, further improving the process.

Lessons Learned

  • Generative AI can add value, but it's important to have a clear, value-added approach and involve the right subject matter experts.
  • Leveraging generative AI can be like having an "intern" - you need to understand the process well to teach the model how to add value.
  • Investing in people with both technical AI and domain-specific knowledge is key to successful implementation.

Continuous Improvement with Reinforcement Learning

  • The team implemented a reinforcement learning approach, where the human decisions on ticket classification feed back into the model, creating a virtuous cycle of continuous improvement.
  • This allows the embedding model used for duplicate detection to be fine-tuned over time, further improving the accuracy and relevance of the results.

Call to Action

  • If you're interested in learning more about how generative AI can improve your ticketing or other business processes, visit the Generative AI Innovation Center in the AWS Expo Center at re:Invent.
  • The team is available to discuss your specific challenges and explore how generative AI can be leveraged to drive tangible business value.

Your Digital Journey deserves a great story.

Build one with us.

Cookies Icon

These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.

If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.

Talk to us