Next-generation prompting for engineering: Accelerate new engineer onboarding, debugging, and code reviews

Neil Dhruva

Engineering

For engineers, proper onboarding, staying up to speed, and understanding how best to proceed based on context is critical to writing and maintaining quality code. Work AI platforms can assist engineering teams achieve those objectives, but it's not always easy to get started with building AI workflows and prompts. 

Our goal here at Glean is to enable everyone to be capable of leveraging great prompts, making it easy for anyone in your organization to find and use AI to boost productivity. To that end, we're thrilled to introduce our next-generation prompting features, designed to revolutionize how you interact with AI in your daily tasks. 

Glean's next-generation prompting offers significant benefits for engineering teams by automating tasks and accelerating usually tedious processes. Users can create multi-step prompts using natural language, eliminating the need for complex prompt engineering or additional coding. By simplifying the process of creating and using prompts, Glean enables engineering teams to streamline their workflows and focus more on high-value activities.

Furthermore, Glean's prompt library enables you to easily discover and share useful prompts across your team, enabling anyone to harness the full potential of AI. In this blog, we’ll explore how NGP helps engineers get up to speed, gather context, and review code quicker than ever.

Onboard, debug, and review with Glean’s prompting capabilities

One of the biggest challenges in any engineering organization is getting new hires up to speed quickly. Our engineering team at Glean has been growing rapidly, and onboarding new engineers efficiently has become a critical focus. 

Thankfully, new engineers have reported that Glean has provided them with the fastest onboarding experience of their careers. This is because Glean not only helps them understand how things work, but also provides context on past decisions and identifies relevant experts within the organization. 

One of the most time-consuming aspects of onboarding is preparing for production on-call duties, which requires the ability to troubleshoot a wide range of issues or knowing exactly who to contact for help. Glean's advanced prompts, multi-step prompts that automate complex analysis and workflows, have been a game-changer for this task. When an engineer is on call, they can use our on call prompt to quickly access key information such as related issues, possible playbooks for resolving the issue, and a list of experts to consult if needed. This reduces the time and effort required to troubleshoot issues, especially for new team members who may lack necessary context.

This on call advanced prompt provides everything engineers need to get started with issue resolution and next steps

Another prompt that’s really impactful is the Glean reviewer prompt, which automates much of the code review process for pull requests. Glean reviewer is an advanced prompt that utilizes multi-step retrieval, which isolates individual steps each with their own retrieval and reasoning, and then links them together to form a complex process. It pulls information from GitHub, Jira, and Slack, then reads and understands code changes, gathers additional context, and provides insights on actionable issues along with a summary of suggested changes by leveraging our internal APIs. It even checks for adherence to internal style guides! This has significantly reduced the time spent on code reviews, cutting down on iteration times by nearly half.

The Glean reviewer prompt makes code review easier, faster, and more comprehensive

Glean's next-generation prompting enables engineering teams to accelerate workflows like code reviews, onboarding, and on call duties—along with many other tasks. Here are a few other prompts that engineering teams have found useful for their daily workflows: 

  • Find relevant guides, documents, or code files to begin writing code

    • I'd like to start writing code for this feature: [[Feature name]]

      First, search for only the feature name itself (without any extra keywords) to find any wikis, documents, or code files that are relevant.

      Then display the ones that would be most helpful for someone getting started on this project. Prioritize getting started guides and documentation, followed by potentially relevant code files. Provide a link to each of these files.
  • Identify the cause of an error message and how to resolve the issue

    • When I run the following code:

      ``` [[Code snippet]] ```

      I am getting the following error message:

      ``` [[Error message]] ```

      Please suggest the cause of the error and how I might fix the code.
  • Identify individuals who have worked on a feature in the past

    • Which people have worked on this feature in the past month:
      [[Feature name]]

      Search for only the feature name (without any extra keywords) and extract the key people who have recently modified the feature.

      For each person, display their name, email address, and a summary of all the changes they made to the feature during this time period.

The future of AI — “beyond the box”

When Glean first launched, our focus was on enterprise search, enabling users to input queries and receive answers within a box. We then expanded to include chat functionalities, integrating conversational interactions within the same confines. But now, we're breaking free from the box, helping everyone expand their potential to do extraordinary work. 

To learn more about Glean’s next-generation prompting capabilities, check out our dedicated blog, the prompt library, and watch the Glean GO event, where you’ll see demos of additional prompts and use cases being put into action in everyday work. If you’re looking to transform your employees’ workflows and business with an enterprise-ready Work AI platform, get a personalized demo today!

Related articles

No items found.