Automotive GenAI sol to better cust care & field eng process

Our voice enabled GenAI Chatbot can cut down Service calls significantly, scale up service dept, technician can be guided with proper response & problem resolution, immediate fix, reducing downtime.

Comment

Inspiration

In automotive industry, automation has been a part of the automotive industry for quite sometime, but still there are few challenges faced by field engineers of automotive industry. The main challenges in the automotive sector are that service process include repairing delays, high customer demands for quality services at low cost and reduced execution times, and high competitiveness among micro and small companies. Most field service organizations operate in saturated markets where the key to success hinges on competitive differentiation and great customer service. There are huge number of resources (owner manuals, process manuals, troubleshot documents, technical documents etc.) provided by the car manufacturing companies for each of the car model promoting in the market. Service engineers has to go through the documents and inspect the issues faced by customers and provide their solution. Inefficiencies in service information transfer, including insufficiently detailed information provided to service designers, incompatible databases, and multiple formats of information for service technicians, contribute to inefficiency in the service process . Slow demand recovery and the lack of inputs are affecting production chains, leading to temporary production halts and challenges in competitiveness.

To address all these we have prepared one Gen AI enabled chatbot solution using Databricks and AWS for car owners and service engineers which will assist then to address any technical issues related to cars and its parts. We prepared the knowledge base by ingesting a number of Car process manuals, owner manuals, service engineers manuals, technical documentations etc. as form of PDF documents and created the Databricks Vector DB. Databricks BGE large embedding model to generate embedding and DBRX to do contextual search on the enterprise data. We have also providing the voice assistance to the customers and service engineers to get the response quickly and easy to understand.

What it does

Our Gen AI enabled chatbot solution can improve this situation by providing below advantages: Service calls cut down significantly – By using the chatbot customers can directly query the configuration and issues faced to the chatbot and get necessary information. Benefit for Service Department – It can scale up service department by handling multiple customers at one time. It make it possible to scale through digital transformation without hiring more employees to perform the same set of tasks. Automated handling of service request - By going online to request technical support, the customer or technician can transition with ease through an automated process that doesn’t require actual human interference until a technician receives the request to visit the customer site. In field service, it can support technicians in the field. Video Assistance Chatbot can be used as a resource to help technicians when they are having issues resolving a service order. The technician can be guided by the bot to proper documentation and problem resolution, or they can request immediate visual support from a remote expert to guide them towards an immediate fix, reducing downtime and the need for additional truck rolls.

How we built it

We have built it using Databricks PySpark notebooks, Databricks Vector DB, Databricks DBRX LLM, Amazon Transcribe for voice conversion, S3 storage, Databricks Legacy model serving, React JS for UI

Challenges we ran into

There are few challenges to hosting the model as we are using Databricks free workspace. So we have enabled legacy serving and hosted our RAG based model in that. As there is a voice assistance, we faced issues to convert voice to text to get the query from user's voice input and text to voice to be provided to the user. We have used Amazon Transcribe batch jobs for voice to text conversion and Amazon Polly to text to voice conversion.

Accomplishments that we're proud of

Our solution received appreciations from TCS leadership team. We have showcased to our customers as well and received appreciations. They liked the solution and the way we have built it.

What we learned

Gen AI can resolve a large number of business problems, customers self service, automation & efficiency in multiple business domains.

What's next for Automotive GenAI sol to better cust care & field eng process

To provide a visual assistant so that service engineers can get visual support from a remote expert to guide them towards an immediate fix, reducing downtime and the need for additional truck rolls.

Built With

Share this project:

Updates