Priorities: How to Evaluate Task Ranking Quality

What the Tool Actually Does

'Priorities' is a text-based tool. You enter three things: a task list (the 'Your text or idea' field), available time (the 'Available time' field), and a goal or desired result (the 'Goal or desired result' field). The output is a single text result in a separate panel: a ranked list with explanations of assumptions and dependencies. The result can be copied or downloaded as a plain TXT file.

It is important to understand what the tool does not do. It does not automatically save results, does not maintain a task database, does not integrate with a calendar or tracker, does not send notifications, and does not publish anything directly. There is no batch processing, no registration, and no premium plan. This is a one-time text analysis, not a project management system.

Therefore, the quality of the result is not evaluated by the beauty of formatting, but by three criteria: ranking accuracy relative to your goal, completeness of explanations (are assumptions and dependencies visible?), and suitability for further work — can you create a real plan for the day or week based on this text?

Example input: input — 'Fix formatting order bug, update logo colors, reply to client, plan content for next month'; time — 'today until 18:00, about 5 hours'; goal — 'don't miss the launch and don't lose the client'.

How to Prepare Input So the Result Is Checkable

The more specific the formulations, the easier it is to evaluate the output. Instead of 'fix a bug', write 'fix formatting order bug, due to which payment falls with an error'. Instead of 'reply to client', write 'reply to client who is waiting for delivery deadline answer'. Formulations must contain a sign of urgency or importance, otherwise the tool has to guess, and you won't be able to distinguish a reasonable assumption from an invention.

In the time field, indicate the real budget, not the desired. 'Today until 18:00, about 5 hours' is good input. 'A lot of time' is bad: the tool cannot correlate tasks with a constraint. If the time is unknown, honestly write 'about 3 hours, I don't know exactly' — this is also useful input.

In the goal field, describe what is more important to you: don't miss the launch, preserve the client, stay within budget. The goal is the main criterion by which you will check the ranking. If the goal is vague ('do everything'), the result will be hard to evaluate, because any ranking can be called valid.

Example input: input — 'Fix formatting order bug, update logo colors, reply to client, plan content for next month'; time — 'today until 18:00, about 5 hours'; goal — 'don't miss the launch and don't lose the client'.

Checking Accuracy: Does the Order Match Your Goal?

The first step of verification is comparing the task order with your goal. In the example goal — 'don't miss the launch and don't lose the client'. Therefore, the formatting order bug blocking payment should be placed higher than updating logo colors. Replying to the client waiting for delivery deadlines should also be high, as it is directly linked to the risk of losing the client. Planning content for the month is important but not urgent, and can be moved down.

If the tool puts logo update above the payment bug, this is a signal of error: either the input was unclear, or the model incorrectly prioritized. Check if you wrote something about the logo being 'urgent, client is waiting', which could skew the ranking.

Second step — check if dependencies are considered. For example, replying to the client may depend on whether the bug is fixed: if delivery deadlines depend on system status, you first need to understand the bug status. A good result explicitly names such connections, rather than just producing a list.

Expected order: 1) formatting order bug, 2) reply to client, 3) update logo colors, 4) content plan. If the logo is above the bug, check the input formulations.

Checking Completeness: Are Assumptions and Dependencies Visible?

The quality of the result largely depends on how transparent the assumptions are. The tool should explain why it considers a task urgent or important. For example: 'I assume the bug blocks payment, therefore it is critical for launch'. Such an explanation can be checked and disputed. If the output just lists tasks without justifications, it is almost impossible to evaluate.

Pay attention to dependencies between tasks. In the example, content planning may depend on which logo colors are approved, if the content uses branding. A good result will note this. A poor one will ignore it and create an illusion that tasks are independent.

Check that no tasks from your list are missed. The tool should not silently discard items. If it combined two tasks into one, that is acceptable, but it must be explicitly stated. Completeness is not only about having all points in place, but also about all significant connections between them.

Expected order: 1) formatting order bug, 2) reply to client, 3) update logo colors, 4) content plan. If the logo is above the bug, check the input formulations.

Checking Suitability: Can You Work from This Result?

Even an accurate and complete result is useless if you cannot make a plan from it. Check if the output has a clear sequence of actions: what to do first, second, and what can be postponed. If the result just says 'the bug is important', but does not help decide whether to tackle it immediately or reply to the client first, it is not specific enough.

Evaluate if the result matches your real time budget. If you have 5 hours a day, and the tool suggests doing all four tasks with equal depth, it is unrealistic. A good result either highlights priorities or honestly states that with such a limit, some tasks must be moved.

Finally, check if the text can be used as a draft for further work. You can copy it into your planner, edit the formulations, and add details. If after reading you still have the question 'what now?', the result needs to be reformulated with clearer input.

Expected order: 1) formatting order bug, 2) reply to client, 3) update logo colors, 4) content plan. If the logo is above the bug, check the input formulations.

Limitations and Typical Errors

The tool works with text and does not know your context beyond the input. It does not verify facts, does not know real deadlines, does not see your workload, and cannot guarantee that prioritization will lead to the desired result. The model can make mistakes, and queue availability may change. This is a draft assistant, not a source of truth.

A typical error — too general formulations. 'Make a report', 'make a call', 'figure out the task' do not give the tool material for ranking, and it starts guessing. Another error — contradictory goal: if you write 'manage everything' and at the same time indicate 2 hours of time, the result will be internally contradictory.

Another error — blind trust. Even a good result should be double-checked: does it correspond to your obligations to clients and colleagues, are there tasks that cannot be moved due to external reasons? The tool does not know about agreements not present in the text.

Remember the input limits: the first field — up to 8000 characters, the details field — up to 1000. This is enough for a task list, but not for a full project description. If there are many tasks, it is better to divide them into logical groups and process them sequentially.

Expected order: 1) formatting order bug, 2) reply to client, 3) update logo colors, 4) content plan. If the logo is above the bug, check the input formulations.

Practical Verification Checklist

After generation, go through five points. First: are all tasks from the input present in the result? Second: does the order correspond to your goal? Third: are the assumptions underlying the ranking named? Fourth: are dependencies between tasks marked? Fifth: can you make a plan based on this text for the available time?

If at least one point is not met, try to clarify the input: add an urgency sign, clarify the goal, reduce the list to the most important tasks. Often the problem is not in the tool, but in the fact that the original formulations allowed too many interpretations.

It is useful to save a successful result in TXT and compare it with what you actually did during the day. This is not a built-in function, but your own habit: it helps understand where the ranking was accurate, and where you overestimated or underestimated a task. Over time you will learn to formulate the input so that the result is closer to reality.

Expected order: 1) formatting order bug, 2) reply to client, 3) update logo colors, 4) content plan. If the logo is above the bug, check the input formulations.