4 ways AI frees me from product management busywork
AI does the busywork so I can do the thinking.
Ayuba Audu
Product Manager at Coda
4 ways AI frees me from product management busywork
By Ayuba Audu
Share this article
AI · 6 min read
In your average day, how much time do you have to spend on repetitive admin? If you’re a product manager like me, I’d guess a good percentage of your working hours are spent on sending requests for updates from your team or generating similar reports for execs—often using exactly the same language over and over again. Coordinating, organizing, and hurrying things along requires careful attention and relatively constant checking in. That’s what I love about working on Coda. Not only has it created a succinct system for my monotonous admin tasks, it’s helped me get valuable time back in my day. And who doesn’t want that? The main thing I’ve found from working on Coda —which I suspect any product manager craves—is freedom. I’m free from the daily admin that gets in the way of actually directing a project, the continual spreadsheet updates, check-in messages, and redrafting of the same report. I’m free to focus on the highest leverage elements of my work, like thinking through (and often preventing!) issues with our strategy—sometimes even using Coda’s new turnkey AI platform, Coda Brain, to help do so. In short: AI does the admin so I can do the thinking. Here are just a few ways AI has helped free my admin time, uplift my productivity, and elevate my day-to-day as a PM.
I hope it goes without saying that I’m a Coda fan. I work on and in Coda every day, sketching out potential new features from within the platform itself. As a PM, that involves talking to dozens of people, both on my team and outside of it, and tracking several different initiatives at the same time. Our team’s doc is an all-in-one productivity space. Project trackers live next to our data, meeting decision logs, PRDs, integrations from our most used apps, and anything else we need, and it’s all automatically updated as we move through our tasks. At Coda, I can centralize everything in our team hub and use AI to take meeting notes, generate to-dos based on those notes, and assign the tasks to my team’s dashboards all on its own. As a self-diagnosed tab hoarder, life is easier when you have fewer tabs. Centralizing our information and syncing it to auto-update saves hours of my time every week. Important side effects include a happier, more cohesive team and execs who don’t have to wonder what’s happening downstream. While my workflow has moved on from spreadsheets, my communication with my team does still live in Slack. If you’re anything like me, you’ve sent some variation of “Hey, can I get an update on XYZ” so many times that now you copy the message from last week, edit, and send. These days, I rarely have to type it out myself. If a task is ever overdue, my dashboard can send an AI-generated response, personalized Slack, and/or an email to follow up on it. I doubt I need to convince you of how helpful it is to shuffle the most basic of admin tasks off my docket.
Automating things like due date reminders actually means I’m communicating better, and our entire team’s expectations are more aligned. My communications are more consistent and accurate, they take less of my time and energy, and my team knows exactly when to expect them. I never miss something when it’s turned in, and they know a follow-up is inevitable. And that’s without any extra effort on my part.
Coda has been around for 10 years in one form or another. While I’ve only been around for 2 of those 10, I want to make sure the next decade is as great as the first. So, as someone who is building on a foundation I didn’t pour, I often need access to data I didn’t generate. I need to know what we’ve tried in the past and how it went, so I’m not wasting the time of my whole team retreading old ground. Since what I often need is proprietary data, consumer AI like ChatGPT (or even more fact-focused alternatives like Perplexity) can’t answer my questions about Coda’s history with any degree of accuracy. They simply don’t have access to that information. Coda Brain does. Coda Brain is kind of like ChatGPT if it could function at an enterprise level, in that it can talk to you and answer questions. But it was also built with Retrieval Augmented Generation (RAG), meaning Coda Brain won’t answer your question based on a likely sounding Reddit post. With RAG, Coda Brain has access to all the data I have access to (and nothing more), whether that data lives in Coda or tools like Jira, Google Calendar, and Slack. So, when I ask it a question, Coda Brain looks through all that data to answer my questions. This means Coda Brain will not only answer questions in text (unstructured) or tables (structured) across all your tools, it’ll always cite its sources.
As we work on new product updates, Coda Brain has saved us from wasting time retracing old ground. With real access to our institutional knowledge, my team doesn’t have to waste frustrating hours figuring out the context of their tasks. We have access to decades of accumulated wisdom, and we can put it to use much more quickly than we could without Coda Brain.
You don’t always get to PM projects that will actually change the way you work, but it’s a delight when you do. Approaching a project as a user and manager helped me think holistically about what I as a user would want Coda to do. As a product manager, my dream tool helps me identify and prioritize the most impactful product features. It’s been exciting to turn that into a reality. Coda Brain can show me all customer requests for any given feature and sort them into a visually appealing, constantly updating table. It can find notes from any previous explorations and brainstorms we’ve made in a specific direction and pull records of decisions. These are features I will use in setting the direction of my team for every new project. One of my favorite results of this interplay between my work and my projects has been Coda Brain’s easy reporting capabilities. Our docs and tables (which function like relational databases) can generate different views for different audiences with a click or two, but Coda Brain can create brand new tables of our current or historical data. That’s more information, presented with better organization and context, generated instantly. That’s hours of research and design work that can now be put to better use.
Generative AI is fun as an art tool, but, given the right access and permissions, it’s powerful as an admin tool. You could generate a limerick about your favorite co-worker just for a fun distraction. Or you could offload the work of summarizing a call and generating to-do lists entirely. The things Coda Brain can do can change the trajectory of your workday, not just entertain. If AI is going to have any lasting impact on the way we work, why not focus on the useful, time-saving admin work? We should be teaching our teams how to use AI well, instead of trying to make it replace graphic designers and writers wholesale. Working on Coda has only deepened my conviction that properly implemented AI will help us eliminate busywork, not creativity. If anything, getting calls and reports off my to-do list has left me with more room to generate new ideas. And yes, Coda Brain helps with those sometimes, too. With access to your data, Coda Brain can be a useful brainstorming tool. I’ve been able to ask it to assess whether or not milestones for a particular project are at risk of happening on time, and the AI has been able to point out a problem or two I missed. I know my sales colleagues have asked for a list of the best potential sales opportunities over $100,000. Coda Brain isn’t replacing anyone on the team, but it is helping them work twice as quickly (and efficiently).
If you’ve been feeling like a professional copy-paste machine lately, I see you. But I hope the promise of AI will help you think about ways you could offload those repetitive admin tasks that take up so much of your time. Take a look at what Coda Brain could do for you today.