How I’m using AI to 10x my work as a product manager

As a PM, AI is saving me hours of time in my day-to-day work. Plus, it’s helping me make better decisions, too.

Ayuba Audu

Product Manager at Coda

How I’m using AI to 10x my work as a product manager

By Ayuba Audu

Share this article

Share on TwitterShare on LinkedInShare on Facebook
Product teams · 7 min read
One thing that’s true for most product managers is that we love solving problems. And one problem most of us PMs have is a lack of time. That’s why, when AI suddenly got good (like, really good) and widely available with the release of ChatGPT, I couldn’t wait to get my hands on it and see how it could help me. Since then, I have fully embraced using AI in both my personal life and at work. AI is now saving me literal hours by cutting down “busywork” and eliminating time-consuming manual tasks. Plus, it’s helping me make better decisions and write better docs by providing me with a virtual thought partner. Here’s how I’m using Coda AI, our workplace assistant, and Coda Brain, our turnkey enterprise AI platform to help me 10x my work as a PM.

1. Prioritizing what to build.

 Any PM will be familiar with the challenges of defining roadmaps while juggling requests from users, your sales team, the CEO, support tickets, and so on (not to mention your own ideas). Prioritizing which features to build is never simple, so at Coda we like to use some “rituals” to help with the process. My favorite is the $100 voting exercise introduced by our CPO, Lane Shackleton. This simple game allows you to capture both what your team thinks should be prioritized and also how strongly they feel about it. That helps identify, for example, that the #1 idea is four times as important as the #2 idea.
Ahead of this exercise, I give AI some context about our overall goals and asked it to suggest a few feature ideas. Nothing matches the creativity that comes from our team discussing and building on ideas as a group, but these AI-generated ones are useful for getting us started. Once we’re done voting, I use AI to summarize any notes from the discussion. I then finalize the list of priorities using the summary and also the additional context I have in my head. For example, I might decide not to prioritize the most voted for idea because I know it’s a bigger scope than we can resource, or there’s something similar already in flight. Or, I might decide to prioritize an idea that’s lower down the list, because I know that three of our biggest customers are asking for it. I then use AI to help me draft a follow-up for everyone who took part. This includes the list of prioritized ideas—let’s say a desktop app being top in this hypothetical scenario—and the next steps. I always refine what the AI has drafted to add my own voice and any additional context as noted above.

2. Defining the scope.

 Since we’ve decided to build a desktop app, we next need to define the vision and scope of the project. I start with user research, but rather than reading through rows and rows of customer feedback, I use AI to consolidate and summarize the main themes. Now I have a useful overview of what’s most important to customers to help define our “North Star.” I also want to see what our competitors are doing, so I use AI to generate a list along with information about their desktop apps and what features they have. Again, this has saved me literally hours of trawling through websites and help documentation, allowing me to move fast and start defining scope sooner rather than later.
I always do a bit of sense checking and refinement of the list. This is for two reasons: Firstly, the datasets AI uses aren’t always up-to-the-minute, so I want to check if there have been any recent updates or launches from these competitors. And secondly, I might want to trim down the list if I know there are certain competitors or features that aren’t as relevant to us. Coda Brain makes this process easy for me by including its sources so I can verify as needed. Once I have this AI-gathered data, I also use my own knowledge—such as the fact that we tried out a similar idea at a recent hackathon—and factor that in to the scoping decisions.

3. Setting goals.

 We always set goals before we begin development on a project. That way, we know what we’re aiming for and can track the project’s success. We want these goals to be tied into the company vision and OKRs, so they’re strategically aligned. I asked AI to gather a few different data points:
  • Which company OKRs are most related to our project (in the case of our hypothetical desktop app, that might be increased usage and customer retention, for example).
  • Goals and results for similar projects in the past, so I can use these as a benchmark.
  • Other ideas for how we might track success of the project in case there’s anything I’ve overlooked. I often use AI in this way to “spark” more ideas or to validate what I already have.
I then use all these inputs—plus my product sense—to finalize our goals for the project.

4. Crafting a PRD.

 Once I have a good idea of our scope and goals, I need to write up a product requirements doc (PRD) to serve as a guide for everyone on what we’re building, why we’re building it, and how we’ll achieve success. I use AI throughout this process. First, I use it to gather and summarize any previous write-ups, explorations, or conversations we’ve had about a desktop app in the past. That way, I can take any previous considerations or blockers into account. Then, I ask AI to draft the PRD using our standard template and the information I’ve collected so far. I further refine it to make sure it aligns with my voice and is in the tone and format that will be most persuasive. This is something AI would struggle to do, because it doesn’t have the in-depth knowledge of exactly who I’m speaking to (i.e., my specific colleagues), how previous PRDs have been received, and what is typically most effective in our specific company.
Once I’ve done that, I turn back to AI to get some feedback on the doc. Coda AI adds comments with suggestions to improve clarity, tone, and grammar. It even adds some praise, too! Sometimes, I ask the AI to adopt a specific persona (say, a product leader at a high growth B2B startup) or ask it to review for certain things, like repetition, to get even more targeted feedback.

5. Kicking off the project.

 The last step before we start building is to set up our project tracker and assign the work to the relevant people. We usually include automations that alert each person when their tasks are ready to be worked on, based on previous tasks being completed. To avoid starting from a completely blank page, I ask AI to analyze my PRD and make a task list that I can use as a jumping off point. I’ll then refine it to get to our complete list.
I use my knowledge about everyone in our team, including each individual’s strengths and interests, to assign tasks to the best people. Then, for our weekly stand-up meetings, I use AI to summarize the notes on each person’s tasks and identify any themes we need to discuss. This helps to keep our weekly team sync quick and efficient.

6. Keeping everyone in the loop.

 Within our task tracker I always include a view that shows progress across the project. This dashboard auto-updates based on the status of the tasks, and largely cuts down on the “how are we progressing?” questions I get from the team and our stakeholders. I also like to send out weekly updates in Slack to keep our key stakeholders up to date. I use AI to draft this message each week, which I then refine and push to Slack using a button in my dashboard.
Getting AI to do the first draft greatly speeds up this process, but I always like to add in some personalization and any additional context. For example, I might call out that it’s a new team member’s first pull request, or include a bit of customer feedback we’ve received. This makes the updates feel more engaging and more likely to be read (which is, after all, the goal!).

AI helps me be a 10x product manager.

 These are just a few examples of the many ways I’m using AI to help in my day-to-day work, while allowing my unique skills to shine throughout the process. I’ve found I can go deeper—using AI as a thought partner to give me leverage—on tasks requiring creativity, navigating relationships, and understanding the broader context. I hope this post has given you some inspiration for how you might use AI in your own work. If you want to try it for yourself, Coda AI is free for all doc makers and can help with a whole range of tasks, like giving feedback, drafting content, finding answers, and summarizing data. Get started with Coda, or learn more about Coda AI and Coda Brain.
Try it for yourself

Related posts

Explore more resources for product teams.
How AI frees me from busywork

AI does the busywork so I can do the thinking.

4 AI predictions

With the rapid advances in AI technology, what’s coming next for us as PMs?

Writing better AI prompts

Tips for crafting effective prompts when you need more than simple answers.