We've all used some form of project management system at work. Here are some common examples:
Asana : used for issue tracking
Smartsheet : used for work scheduling
AirTable: used for generic project work
Trello: used to manage work items in a Kanban style
Jira: used to manage software bugs
Intercom: used to manage customer support tickets
A structured spreadsheet: used to keep track of anything
They have minor differences, but if you've used one of them, you'll be reasonably familiar with how all of them work. While they assert that they enable collaboration, the actual work gets done by people outside these systems. The project management systems are esentially a database to record work and to report on progress. Important of course, but not intelligent. They help productivity to some extent by avoiding confusion, but they don't actually do any work.
It is important to provide the core capabilities that users have come to expect of these systems. Thunk.AI certainly does that. But it also adds a lot of new value because it is AI-first.
How does an AI-first design change a project system?
There is a huge difference between an AI-first project environment and one of these legacy project management systems. Just as by analogy, there is a huge difference between having a colleague who co-authors an article for you and having a piece of paper to write on.
Planning: Thunk.AI uses AI to actually plan the work tasks to do. On the other hand, when legacy systems refer to planning, they usually mean "scheduling" work and allocating resources.
Automation: Thunk.AI uses AI to drive the automatic generation and execution of tasks. In most legacy systems, it is upto humans to create tasks manually and record them in the system.
Task execution: None of the legacy project systems attempt to understand the actual tasks to be done, let alone take on the work of executing them. They are not intended to be AI agents.
Data: like any project system, Thunk.AI can represent structured data. But uniquely, Thunk.AI deduces the data structures needed for the problem at hand. It is AI that steers the construction of the data.
Collaboration: many legacy systems enable collaboration via a shared editing user experience and a lot of noisy messages. Thunk.AI takes the noise out of collaboration because AI agents process most messages and do most of the data editing.
Workflow: In most project systems, workflow is somewhat of an appendage, implemented through mechanisms like trigger scripts that kick of a rigid sequence of actions. In Thunk.AI, process workflow lies at the heart of each thunk. It is defined in natural language, guided by the knowledge of the AI model, and is intelligent in adapting to each specific work item.
Messaging: Most legacy project systems send a lot of messages to users to keep them abreast of changes in project data. The most common message is: "You have been assigned workitem X". On the contrary, in Thunk.AI, the user's AI agent handles messages and does most of the work. So it is much more likely for the user to see: "Good morning, while you were asleep, I've received and handled these N workitems for you".
Reporting: Legacy project systems focus on the needs of the project manager. It's all about making sure the team members update their status in the system so that the reports are accurate. Thunk.AI also has similar mechanisms like Kanban boards to visualize project work status. Yet, the key difference is that the work is being done by the AI agents, they are updating work item status, and therefore the project reports are far more uptodate without a lot of wasteful human bookkeeping.