Fulfilling the potential of task-management apps

AI creates a unique opportunity for task-management software to become what it always dreamt of being. And the reverse is also true: task-management software is a unique opportunity for AI output to be a lot closer to what we need to reach our goals.

When working on its own, AI constantly goes from planning mode to doing the work. Planning and doing feed into each other. It’s a marvel to see AI working on a complicated goal, constantly adjusting phases and tasks, reviewing and iterating again. AI, with task-management software, gives us the opportunity to work using a similarly tight plan-work-plan loop.

We already tend to work in a plan-work loop, using AI or not, in one way or another. In my case, it is quite messy though. Well, it used to be anyway. I used to have a bit of it in a todo app, maybe a spreadsheet, a bit in my head, another bit in a note-taking app, and maybe some on paper. And I alternate between these in the journey that is reaching my goal. If there is no incentive to put updates and notes in one place, I won’t. And of course, the issue is that I’ll lose focus on the original target. And I’ll lose the opportunity to reward myself, and set new goals.

I am an engineer. What I’m describing above, the broken “plan-work loop”, is a lot more foundational and universal than the terms “plan”, “work”, “loop” themselves. If I were to put a bit more humanity into it, I’d use “decide” instead of “plan”, “do” instead of “work”, and “take a step back” instead of “loop”. So decide, do, and take a step back. Following these steps consistently is of course critical to reaching our goals, and get that feeling of accomplishment.

So, how can AI help with making our lives feel more accomplished?

1) Helping us to get started. AI can draw from its close to infinite (and accurate-ish) knowledge to get an idea of how to reach any goal, and give a starting plan. 2) Another one is AI helping to figure out difficult tasks. That’s what we tend to already use AI for. For example, what are good options for a flower bed in the shade? AI knows. 3) Next one is feedback. We all crave feedback, and AI is happy to deliver. 4) Yet another critical one is coaching. What is the next important thing to work on? Maybe a little pep talk could also help with starting the day. 5) By always having a plan. Updating the plan to match reality is as valuable as it is difficult. For AI, adjusting plans is a core capability.

There is an obvious one I am leaving out of the list, and that is AI performing the task itself. It can help with the feeling of accomplishment by making possible previously unreachable goals. It can also rob us of a satisfying struggle to accomplish something difficult.

Now let’s focus on why and how the task-management app model can be the perfect tool to enable this human-AI collaboration.

Imagine your todo app embeds your AI. You click on the task and you get help. And that help comes with the knowledge of the task, and the overall context of what you’re trying to achieve, and any intermediate results, notes or photos. AI’s response quality is as good as its context, and having access to the full goal picture is hard to beat.

Let’s push this a little further. One evolution of AI chat nowadays is to become more than a chat and become workspaces. We could even generalise that observation and mention that software as a whole is becoming a workspace for human-AI collaboration. Coming back to the idea of working within a task, this means the task you clicked on can become that workspace, specialised for that task. In practice this can mean AI skills, or even UIs and functionalities targeted at the task at hand.

You might ask, what about specialised software? Specialised software can be powerful and irreplaceable for some tasks, but that power comes with complexity and takes skill to use. For the tasks needing specialised software, the task workspace can be the glue between the app’s capabilities and the rest of your project. Today, general purpose AI chats are going in that direction. But I argue the task-management app is a better fit for that glue: the task already feeds AI with what we’re trying to do, and what was done before. AI already has the context of the overall goal, previous results, notes and any supporting evidence, and what needs to be done.

The change of perspective is the task-management app no longer being just a record of work to do. Our task-management app becomes an active plan that’s directly helping to do the work. And what we learn while doing the work feeds back into the plan: what is finished, what is blocked, and what should happen next. For tasks in the digital world, the work can be started, produced, and reviewed within the task itself. Where could this approach help? Everywhere. Think graphic design, finance, coding, writing. And for the non-digital world, smartphone cameras and wearable tech are the bridges to unlock similar benefits.

“But I don’t want to turn everything in my life into a task!” - me

I definitely recognise myself there :). To be fair, I’ve changed a little bit. I’ve always known it, but age finally made me internalise that to do non-trivial things, I need to split them into manageable chunks and properly turn my attention to them. It turns out I have a lot of these non-trivial things in my life: DIY, gardening, admin, fitness, learning, travelling, hobbies, or even party organising. Not everything needs a plan, of course. But if we want the goal to matter, we need a plan.

Note about Task Matter: the above analysis is what motivated the Task Matter project. Theoretically, the above approach sounds great, but does it work? The Task Matter project’s objective is to implement it, and test it. Today, the implementation is only partially there, as the task workspace is fairly rudimentary. And the planning features are bare-bones. But I love it. In particular, the Task Matter implementation plan has evolved massively since it started, and the plan still accurately describes where I’m at, and what I want to do. I’ve used it to plan my running goal over 4 months, and I’ve updated that goal based on my fitness multiple times. I’ve used it to help with my application for Irish naturalisation, and it extracted the document requirements for my particular case, with step-by-step instructions. I’ve also used it to plan my DIY projects and a trip to Oslo. And of course, I’ve used Task Matter to build Task Matter, the largest goal I’ve ever worked on.