Case Studies

DRI Copilot: Leverage Generative AI for Support Teams

DRI Copilot by Akvelon
  • Save approximately 40 hrs per month for a support team of 5-6 members.

  • Easily search for similar or duplicate issues and feature requests with Duplicate search.

  • Streamline code snippet analysis to exclude syntax errors with Syntax check.

How Does DRI Copilot Work?

DRI Copilot is a solution based on OpenAI’s GPT-3.5 that uses the power of AI to enable support teams to resolve issues faster and spend less time finding similar cases in the knowledge base. This tool uses generative AI and features Slack integration so that support teams can use one channel for all types of communications.

With the help of DRI Copilot, support teams can:

  • Conduct searches for duplicate issues within specified repositories
  • Estimate priority levels of tickets
  • Identify errors in code syntax within reported issues
  • Generate responses to commonly-encountered issues

DRI Copilot uses GPT-3.5 for data preparation and prediction generation, all while machine learning techniques and statistical analysis methods are used to enhance the search and syntax error identification.

 

The Challenges

Our client has multiple GitHub repositories and uses GitHub’s functionality to ensure that they can address the issues reported by their software users. Also, their support team often forms their backlog by taking the feature requests into work from GitHub.

White GitHub is a great solution for version control and collaboration, using it for product support purposes can be complicated since there are no options to streamline the processes. When a new issue is submitted, there is a chance that such a question was already asked by someone else, so the technical support team often has to search through their knowledge base or look through the replies manually to find the correct answer.

Another challenge that comes with supporting GitHub repositories is the need to check the code snippets that are provided in issues. If the code has syntax errors, the issue can be resolved right away because the bug isn’t on the product’s side.

More importantly, to protect sensitive data, sometimes only one support team member can know the specifics of the product, so some issues can only be resolved by them. However, a DRI has limited working hours capacity and other tasks to work on, so solving some of these issues can take so much time that the requests can sometimes pile on for days.

The main goals of DRI Copilot tool

 

Business Impact

The implementation of our AI-powered tech support assistant has yielded significant benefits for our client’s team, including:

  • Saved 20 hours per month of work for a support team of 5-6 members solely through its Duplicate Search feature.
  • Reduced costs through increased productivity resulting from saved time.

Overall, we have calculated that leveraging AI for technical support saves up to 40 hours per month, leading to improved efficiency and potential cost reductions in product support.

 

 


 

Have a question?

How can AI be used to boost a support team’s capacity?

Utilizing AI for support can significantly enhance a team’s capacity by automating repetitive tasks and streamlining workflows. For instance, AI-powered tools like DRI Copilot can assist support teams by efficiently handling tasks such as duplicate issue detection, priority estimation, and first-line response generation. By offloading these routine tasks to AI, support specialists can focus their expertise on more complex and high-impact issues, ultimately boosting the team’s overall capacity to handle a larger volume of support requests effectively.

What support tools are already using Generative AI for support?

Generative AI is increasingly being integrated into support tools to enhance efficiency and accuracy in issue resolution. DRI Copilot is an example of such a tool, utilizing Generative AI capabilities to generate responses to common GitHub issues. By training models like GPT-3.5 on resolved tickets, our tool can provide timely and relevant solutions to support queries, thereby improving customer satisfaction and streamlining support operations.

Is AI for support teams effective for technical issues resolution?

Yes, AI-powered support tools have proven to be highly effective for resolving technical issues. By leveraging advanced machine learning algorithms and natural language processing capabilities, these tools can quickly identify duplicate issues, analyze code snippets for errors, and generate accurate responses to common technical queries. For example, DRI Copilot’s ability to detect duplicate issues and provide first-line responses based on trained models significantly enhances the efficiency and effectiveness of support teams in resolving technical issues, ultimately leading to improved customer satisfaction and reduced resolution times.