Estimated reading time: 13 minutes

Automating tasks can not only save you time but also help you preserve energy for things you actually want to do.

The new year is a great opportunity to reflect on improvements you can make to your day-to-day work. It’s easy to get so caught up with tight deadlines that you don’t assess how your processes can be improved. But taking a moment to step back, you might see that lightweight automations can save you headaches many times over.

How to identify good candidates for automating

There are a couple of smells you might notice in your daily work that indicate a task is worth automating:

• Does a task involve data entry?
• Are you moving data from one tool to another?
• Do you find yourself using copy-paste?
• Is the task repetitive?
• How many times have you done this task?
• How frequently do you expect to do this task in the future?

If the answer to any of these is “yes,” you can likely set up an automation to help. Frequent tasks are great candidates for automating, but so are infrequent ones. It can be far more energizing to set up an automation than to do a mundane task manually (even if the opportunity is rare and it takes the same amount of time).

Will Larson explains how people “accomplish more if you let them do some energizing work, even if some of that work itself isn’t very important.” Repetitive tasks are rarely energizing, but automating them can be because setting up automations is creative and involves problem-solving. It’s also satisfying to know that you’ve saved yourself time.

Setting up automations

As an engineering leader, you can likely figure out how to code up an automation from scratch, though it’d probably take more time than you’re ready to put in. An integration platform as a service (iPaaS) solution can help here.

Writing effective AI prompts

One of the trickiest things is getting the AI prompt to give you the right answers. Here are a few things that help:

• Set the context for how the AI should act: for example, “You’re a software engineer tasked with grooming the backlog” or “You’re an IT administration assistant tasked with doing a security audit.” This ensures the AI is ready to answer questions with the right context.
• Provide examples of desired output for a given input: to make sure the AI completes the task in the way you’d like, supply some criteria. Write out examples that cover expected and edge cases, supplying precise guidance on how you’d like the output to be formatted.
  As you test your model and it gets things wrong, refine your prompt by including those errors as examples. 
• Break up large AI steps into smaller ones: you may find that some prompts return the correct result while others do not. Break the task into smaller steps and/or multiple prompts to improve accuracy.

Human in the loop

A human-in-the-loop means that there’s a step in the workflow where a person makes a decision or provides input before the workflow proceeds. Jacob Bank, CEO of Relay.app, explains why Human-in-the-Loop steps are so important, “AI can’t do everything. We still have a lot of judgment and context that isn’t available to AI, so while AI can often create great rough drafts, we still need a human touch to create a finished product.”

Including human-in-the-loop steps eliminates worry about any negative impacts since you’re the one making the final call.

Common automations for engineering managers

Every engineer, manager, and team has unique workflows, and one of the benefits of setting up your own automations is that you can fit them to exactly how you work. That being said, here are some common ways that engineering managers and their teams can use automations to improve their daily work.

Ticket management

There are several automations that you can set up to make ticket management more streamlined.

Address tasks stuck in progress for too long

The automation can work like this:

• Once each day, create a summary of all tickets that have been in progress for three days or more. Include links to each ticket and how long they’ve been in progress.
• A prompt for this might look like: “You’re a project manager tasked with summarizing items that need attention from a list of tasks.
  Extract the tasks that have been in progress for three days or more and output them as a list where each task has the task description, task ID, task assignee, and number of days it’s been in progress.”
• Prompt the AI to summarize the list of tasks. 
• Send this summary to yourself via email or Slack message.

Remove stale backlog items (with approval)

The automation can work like this:

• Once each week, have the automation grab all tickets that have been in the backlog for two months or longer.
• Send yourself a Slack message with a list of all of these tickets and the ability to uncheck any tickets that you’d like to keep in the backlog. This incorporates a human-in-the-loop step, which requires your approval.
• All tickets that remain checked will automatically get deleted from the backlog.

Automatically file bug reports from customer support emails

Many of the best software teams fix bugs on the spot instead of managing them. Regardless of how your team approaches bugs, one of the most important things is to have bugs filed in one place.

By making it easy to create quality bug reports from email, you can ensure that bugs all make it into your ticket tracker.

The automation can work like this:

• When a customer sends an email that describes a bug (e.g., a customer bug report or automated system alert), create a ticket via your tracker tool using the body of the email as a description of the issue. 
• Set up a filter to identify emails mentioning keywords like “bug.”
• Use AI to analyze the email and determine the type of bug being reported. 
• Optionally, send a Slack message to the backlog maintainer to get approval to create a ticket with this summary.
• If approved, create a ticket with this bug.
• If your ticket tracker has a native integration with Slack (e.g., Linear, Jira), enable the integration so you can automatically save messages as tickets.

Automations for calendar and email management

As a manager or engineering leader, it’s easy to have your calendar filled up with meetings. AI can help you show up to meetings prepared and send timely follow-ups.

Email yourself a summary of your next day’s meetings

The automation can work like this:

• At 4:30 pm the day before, use an iPaaS provider to summarize all meetings in tomorrow’s schedule.
• For any meetings that have attachments, prompt the AI to extract a summary and any action items.
• Email yourself a list of meetings with the summarized attachments.

Automatically draft follow-up emails for meetings

You likely don’t want to send a follow-up for every meeting in your calendar, but there are certain meeting types – like customer conversations and meetings with external providers – where it’s necessary.

The automation can work like this:

• At 4:30 pm, pull together your meetings for that day that include external attendees.
• Based on the attendees, calendar event name, meeting notes attached to the document, and call recordings, if available, automate a draft follow-up email with likely action items.
• At your convenience, review the follow-up email and make any edits before sending it.

Automatically draft email responses 

There are certain emails, like bug report filings, that you always want to respond to promptly. Using AI, you can automatically draft an email response so that your response is more timely.

The automation can work like this:

• When you receive an email, if the subject line has the prefix “bug,” it triggers the start of the automation. (You can typically do this using a “filter” in your automation). 
• Then, the automation starts drafting an email response that thanks them for filing the bug report and notifies them that the team is looking into it.
• At the same time, the automation files a ticket for the bug.

Final thoughts

There are likely several quick automations that you can set up to save time and energy from repetitive tasks. The next time you find yourself repeatedly doing the same task or doing manual data entry, take a moment to think about if you can save time with an automation!

Get articles like this in your inbox

Choose your LeadDev newsletters to subscribe to.