SikaBot #9

A solution designed to make the communication easier for all Sika users. It provides the needed information quickly by analyzing Sika data in a efficient way.

Comment

Inspiration

Communication issues are something we all face, especially when it comes to communicating with people from different background, interests, majors, .. Solving this and making the customers' lives much easier, and providing a solution that would meet their needs in an efficient way, will make a great impact. Also, helping in controlling the long coming disaster from the top made us inspired to drive full speed ahead.

What it does

Our SikaBot makes communication easier for all Sika users. For example, it gives the user the needed information quickly by analyzing Sika data in an efficient way. As a special feature of the Sika-Knowledge-Hub, the user can upload an image for their problem and SikaBot will reply back with the corresponding information and the way forward for solving the problem.

Features

  • The Backend crawls through the PDFs and extracts text and the images
  • With Azure the text is summarized and tags for the images are detected
  • Users can manually upload PDFs and optionally specify languages in which they want the calculated summary of the PDF to be translated
  • Users can search through all of Sika's PDF documents and images
  • They can then specify one document (set its Id as HubDocumentId and generate a new version (through specifying language and length of new document (keywords, summarization or full text). The users can also specify which pictures they want to have in the newly generated PDF.
  • Through the Report endpoint an Excel file can be retrieved that says how many time which document has ben searched for and how many derivations have been created
  • Through the Backend new chats can be created (targeted towards the Client). To these chats messages can be added which are then answered by the Backend through calling OpenAI.

How we built it

We divided the group work into business insights, architectural design, front-end and back-end. The whole application is hosted on Microsoft Azure. Cognitive Services, Blob Storage, Key Value and App Services are used. The backend code is written with ASP.NET Core in C#.

Challenges we ran into

Correctly integrating Azure Blob Storage was an obstacle for us. Deleting unused Blobs while preserving the integrity of the container was a challenge.

Accomplishments that we're proud of

We had a lot of fun as a team and were able to create a chatbot that works in less than 30h.

What we learned

How to use the new and exciting Azure AI service possibilities. More knowledge about what the company of Sika is doing and their impact on the world and environment.

What's next for SikaBot

Integrating SikaBot to SikaWorld so, it can be a unified SikaHub.

Built With

  • .net
  • android
  • android-studio
  • azurecognitiveservicesforlanguage
  • azurecomputervision
  • azuredatastorage
  • azurekeyvaults
  • azurelanguagetranslatorservice
  • coroutines
  • kotlin
  • openai
Share this project:

Updates