Why Asana Tasks Can't Capture the Thinking Behind Them
Asana's premise: work is tasks. Goals, portfolios, and status tools have been added above them, but the task remains the central unit. Everything else organizes, reports on, or automates it.
That model breaks down the moment someone asks "Which goals are at risk if this vendor falls through?" or "Why did we deprioritize the mobile launch?" The answers live in comments, linked docs, or someone's head. People end up doing the work the tool should do: gathering context, synthesizing it, routing it to whoever needs it next.
Higher-level features, same underlying model
Asana calls its data model the Work Graph. You can link goals to projects, add reference fields, connect portfolios – but the links don't describe the relationship. A risk connects to a goal the same way a task does. The system can't distinguish "achieves" from "impacts."
Goals track what supports them, not what trade-offs were made to get there. Decisions live in task comments, disconnected from the work they shaped. Dependencies are task-to-task, not concept-to-concept.
Asana's AI features inherit these limits. They can summarize activity and flag overdue items, but they can't surface connections the data model doesn't hold. When a deadline slips, the AI can report it. It can't tell you which decision created the dependency or which goals are now exposed. Better AI doesn't overcome the limits of the data it reasons over.
"A system is never the sum of its parts; it's the product of their interaction." Russell L. Ackoff
What changes when context is modeled, not just managed
In Bother, concepts like decisions, risks, and goals are distinct objects in a knowledge graph, connected by relationships the system understands – including dependencies. Open any one and the surrounding context is immediate: what it connects to, what threatens it, what it depends on. The structure of your project is something you see, not something you reconstruct from scattered tasks and threads.
An AI agent keeps that structure current. It proposes connections you haven't drawn, surfaces risks you haven't linked to the work they endanger, flags dependencies that have gone stale. You review each proposal and approve or reject it. The graph stays accurate because maintaining context is a byproduct of using Bother, not a separate chore.
What this means in practice
Evaluating an analytics vendor? Create a task, add subtasks for each option, approximate a scoring matrix with custom fields. The structure is improvised. The rationale ends up in a comment or linked doc, disconnected from the evaluation itself.
Open the decision and the vendor options are arranged around it, each scored against shared criteria in a color-coded matrix – red to green. Select any option to see what it trades off against the others. The rationale – what was weighed, what was ruled out, and why – is part of the evaluation, not something you reconstruct from a comment thread.
Managers flag projects as On Track, At Risk, or Off Track. Smart Status can draft an update from task data – overdue counts, completion percentages, blocked items – but the assessment still depends on someone knowing which risks are active and which dependencies are blocked. The AI summarizes what's in the system; the judgment call stays with whoever has the context.
"Two high-impact risks are unmitigated, and a blocking dependency has stalled for 12 days." The AI generates that from the graph – where risks and dependencies are connected to the work they affect. The task list said everything was on track. The graph shows it isn't.
A new product manager joins. They get access to boards, timelines, and hundreds of completed tasks. The decisions that shaped the roadmap, the risks that were accepted, the trade-offs behind each priority call – none of that is in the tasks. Onboarding means a week of context dumps from colleagues who happen to remember.
They open a goal and see what surrounds it: the risks it carries, the decisions that shaped it, the dependencies it rests on. From any starting point they can follow the thread – a risk to the work it endangers, a decision to the rationale behind it. Every "why did we do it this way?" has an answer attached to the work itself. They never have to ask.
The layer Asana doesn't reach
Asana handles task assignment, deadlines, and workflows. As an Asana alternative, Bother captures the layer those workflows leave out – the decisions, trade-offs, and dependencies that make the work make sense. Organizations pay people to gather and route that context. Bother makes the system carry it instead.
Asana gives you tasks. Bother gives you context.