3 ways to use generative AI for engineering management

There are a plethora of ways to use generative AI tools to streamline and improve on your non-coding, management tasks.

Artificial intelligence (AI) is revolutionizing the software industry. And while there are ample risks and concerns attached to these tools, there are also ways in which they could benefit managers with their daily tasks. 

1. Writing, proposals, and documentation 

As an engineering manager, I write a lot of documents. Writing is an important part of reaching my goals. I must compose documentation and notes to facilitate effective meetings, establish processes and guidelines for my team, and write different proposals. As I frequently write in a language that is not my native one, I can sometimes make mistakes. Consequently, I use tools to help catch these errors and enhance my writing. 

ChatGPT

ChatGPT is an AI chatbot that uses natural language processing to engage in user conversations. Ask it for assistance with various tasks and you will receive an instant response.

It is an excellent tool for writing, as it can help with: 

• Translations
• Grammar
• Readability 
• Tone 
• Audience type 
• Style

Some of the prompts that I use to help me write are the following: 

• Can you check the following paragraph for grammar mistakes and provide a revised paragraph “{paragraph}”? 
• Can you explain what grammar mistakes you have corrected, including why you’ve done so and the grammatical rule in question?
• Can you provide a report on the paragraph? The report should include vocabulary statistics, readability score, and tone type (available options are formal, informal, optimistic, worried, friendly, curious, assertive, encouraging, surprised, or cooperative

2. Performance and process automation 

In addition to writing assistance, generative AI tools can be used in other parts of my work. 

Part of my responsibilities includes managing my team’s performance and reviewing processes to ensure they’re always working at their most efficient. Before AI, I used the following tools to track the essential metrics of my team and to reduce time and workload with automation:

• Jira automation: For categorizing and calculating the priority of tasks, notifying stacked tasks, and adding information to tickets. 
• Slack workflows: Notifying bug states and reminders (feedback, roles, demos).  
• API integrations: To get data and create reports. 

Now, with the help of AI, I have additional instruments to lean on. 

ChatGPT 

ChatGPT can help automate processes with tools you may not be familiar with. 

• Scripts: You can use ChatGPT or AI code tools like Copilot and Tabnine to create scripts to retrieve data using Github and Jira APIs. As an engineering manager, this helps me a lot in analyzing CSVs extracted from Jira, normalizing the data to generate reports, and calculating DORA metrics, for instance.
• Proof of concepts (POCs): When creating POCs, ChatGPT and Copilot can be handy tools that are great to be more efficient in that process. They assist in researching new technologies and tools and evaluating their pros and cons. This tool brings a game-changing advantage by rapidly evaluating the feasibility of an idea.

OpenAI API 

Though ChatGPT is a great tool for supporting you in your tasks, the OpenAI API gives you considerably more freedom. 

As software engineers, we’re familiar with the benefits and adaptability of APIs, alongside their ability to integrate with various processes and tools. I’ve experimented with the OpenAI API in several use cases, including:

• Jira action for enhanced user stories: User stories are general explanations of a software feature written from the perspective of the end-user or customer, and they can be challenging to craft. By integrating Jira automation with the OpenAI API, it’s possible to send a task’s description to OpenAI and receive more accurate and detailed stories in return. Furthermore, the AI can offer suggestions like common user scenarios, acceptance criteria, and feedback. With these tools at our disposal, we can automate and streamline this process, thereby saving time and enhancing our product.
• Summarization: Using LangChain, you can create a tool for summarizing and asking questions about your documents. This is invaluable when quick context is needed.
• Brainstorming: Leverage OpenAI to produce and validate ideas, recommend acceptance tests, and pinpoint potential issues in specific features. Brainstory is a recently launched tool that uses AI to elevate this approach to the next level.
• Onboardings: Design chatbots to streamline the onboarding process, thereby enhancing team efficiency. By using tools like LangChain or Llama 2, among others, you can create your own chatbot with personalized data. This presents a fresh approach to improving the onboarding experience for team members. Instead of merely reading documents, we can now interact with them.

3. Hiring and interviews 

Part of being an engineering manager means participating in the hiring process when expanding your team. This means conducting interviews for a myriad of positions, including back-end engineers, front-end engineers, tech leads, and even other engineering managers. 

Your role in this process is to assess whether a candidate is a good fit for the position. This involves evaluating various skills, including technology, ownership, autonomy, teamwork, process, communication, and impact. However, there are many facets that can make this challenging. For instance, a limited time frame, interviewing someone outside of our expertise, ensuring that the interview process is unbiased, and that the candidate feels at ease to showcase their best self. 

Before LLMs, we solved this by creating specific tests and questions to evaluate different skills. We collated different documents with questions for the different roles, which was, and remains, an excellent approach to conducting good hiring interviews. 

But with the aid of LLMs, we can take it one step further. 

• ChatGPT is a great tool for creating or adapting questions and scenarios in preparation for an interview. It can help you identify the “right” answers for questions about technologies in which you are not an expert, also providing supplementary questions to go deeper. 
• LangChain is great for summarizing resumes or getting questions adapted to a candidate’s CV. 
• Grammarly Go is helpful for cleaning up notes. 

Bias in LLMs 

One of the essential parts of hiring is to ensure that the decision process is as unbiased as possible. Creating specific tests and questions to evaluate different skills helps us go in that direction. AI is suitable for that process, but be careful with delegating decisions. AI can produce degenerate and biased output. For instance, if a company employs AI as a preliminary filter to evaluate candidates, the system might weight them improperly based on its training data, potentially underestimating or overestimating candidates due to their knowledge of a specific language or previous employment at a particular company.

Use the tool to help you, but hiring is one of the most crucial processes in a company, and the onus lies with you to find a candidate who is a good fit.

Balancing utility with privacy

AI tools offer innumerable benefits for our work, but it’s crucial not to overlook privacy considerations. Given AI’s nascent stage, we’re still charting its course, and many aspects will be unveiled with time.

To mitigate any issues in this area, always prioritize data protection when using tools like ChatGPT, and make sure to do so in alignment with your company’s guidelines. 

ChatGPT Enterprise may be helpful in this capacity, which was recently launched to offer businesses a secure way to retain data control. Similarly, Azure OpenAI provides solutions with data confidentiality at their core.

Finding your tools 

While these tools are powerful and can automate a significant portion of repetitive tasks, they can’t replace the unique human skills that an engineering manager brings, such as empathy, critical decision-making, leadership, and the ability to understand and act on complex, ambiguous situations. 

However, as engineering managers, our ultimate goal is to get better outcomes from our teams. If generative AI tools can help us do that by freeing up more time for critical tasks, improving the clarity of our communication, or providing insights we might have otherwise missed, it would be a lost opportunity not to leverage their potential. 

Our jobs will continue to evolve with the advent of AI. The best course of action is to embrace these changes, learn how these tools can enhance our productivity, and utilize them to the best of our abilities. 

Doing this will help us stay updated with the fast-paced technological changes and make us better leaders.