«Find the Gap»: How to Refine Input and Get Precise Analysis
The «Find the Gap» tool helps teachers or mentors see exactly where a learner misunderstood a topic and turn that into specific exercises. The first generated result is almost always a draft: it relies only on the answers you entered in the «Source material» field and your goal formulation. The purpose of this article is to show a concrete iteration: what to refine in the input data, how to recheck the tool's conclusions, and how to bring the list of gaps to a working lesson plan. We do not assign the service functions it does not have: it is a text web tool with two fields and one result in a separate panel, which can be copied or downloaded as a plain TXT.
Open tool: Knowledge gap finder ↗What the Tool Does and What It Receives as Input
"Find the Gap" is a text tool: you fill the «Source material» field (learner answers, fragments of their reasoning, completed steps) and the «Goal or desired result» field (which skill or understanding to check). The limit for the first field is up to 8000 characters, the second up to 1000 characters. After pressing the generate button you get one result in a separate panel; it can be copied or downloaded as a simple TXT file.
It is important to understand the limitation: the tool does not automatically check answers against a hidden database and does not know the context of your course. It analyzes only the text you provided. If the «Source material» contains only correct answers, it will not find gaps — and that is normal, not an error. If you inserted reasoning with an error, the tool can point out the incorrect model and suggest what to train.
Therefore the first result is a hypothesis about misunderstanding, not a diagnosis. Your task on the iteration — either confirm this hypothesis with additional data, or refine the input so the conclusion becomes more accurate. Below is a step-by-step breakdown on a learning example.
«Source material» field: "I think percentages are added directly: a 20% discount, then another 10% means a total of 30%." «Goal» field: "Understand how percentages are sequentially applied to one quantity."
Step 1: Collect Raw Answers, Not Summaries
The most common reason for a poor result is inserting not the learner's answers, but your summary: "he is confused about percentages." The tool has nothing to analyze besides your interpretation. Insert verbatim fragments: how the learner explained the solution, which steps they wrote, where they stopped. The closer the text is to real speech or a recording, the more useful the analysis.
If there are many answers, select 3–5 indicative fragments that relate to one goal. Do not mix different topics in one run: the result will become blurred. It is better to make two or three separate runs with a narrow goal than one general one.
Check that there is at least one error or doubt in the text. If all answers are correct, the tool may only confirm understanding — this is also a useful, but different result.
Instead of "confused about percentages" insert: "20% of 1000 is 200, then 10% is 100, total discount 300, price 700." This shows exactly where the logic breaks.
Step 2: Formulate the Goal as a Checkable Skill
The «Goal or desired result» field sets the analysis frame. A formulation like "I want him to understand percentages" is too wide. The tool works better with a checkable skill: "The learner must be able to calculate the final price after two sequential discounts and explain why percentages do not add up."
Indicate the level: does the learner need to simply calculate, explain in words, or transfer the rule to a new situation? This determines what gaps the tool will highlight: computational error, incorrect model, or inability to explain.
Do not add to the goal requirements that are not in the source material. If you ask to check a skill but did not provide answers about it, the result will rely on guesses. Better narrow the goal down to what is really visible in the text.
Goal: "The learner explains in their own words why sequential discounts of 20% and 10% are not equal to 30%, and gives a numerical example."
Step 3: Read the Result as a List of Hypotheses
The first result usually contains formulations like "possible misunderstanding", "worth checking". Treat them as hypotheses: choose one or two that sound plausible, and check them in the next lesson with an oral question or a short task. Do not accept the whole list at once — this will lead to overload.
Mark which fragment of the source text each point relies on. If the tool points to a gap that is not in your text, it means the goal formulation was too general or you accidentally mixed topics. This is a signal to rephrase the input, not to trust the conclusion.
It is useful to write down the hypotheses in a column and next to each one put a diagnostic task you will use. Thus the draft result turns into a diagnosis plan.
Hypothesis: "The learner adds percentages because they perceive them as independent parts of the whole." Check: task "Price 800, discount 25%, then another 10% — find the final price and explain the steps."
Step 4: Refine Input and Run Again
After checking hypotheses return to the fields. Add new learner answers in the «Source material» field — those you received during the check. Refine the goal if it turned out wider or narrower than needed. Then run generation again and compare two results.
Comparison is the key iteration technique. If both versions point to the same gap, confidence grows. If conclusions diverge, look for the reason in the changed text: perhaps new answers showed that the problem was not in the model, but in arithmetic.
Do not strive for an ideal result on the first attempt. Two or three refinements are usually enough to get a specific list of skills for training. Each run takes minutes, and the benefit of refined input is noticeably higher.
Second run: add the learner's answer after a hint — "25% of 800 is 200, remained 600, 10% of 600 is 60, final 540." Check if the list of gaps changes.
Step 5: Turn Gaps into Targeted Practice
A good result ends not with a statement, but with a suggestion of what to train. From the list of hypotheses select 2–3 skills and come up with a short exercise for each: a numerical example, an oral explanation, finding an error in someone else's solution. Practice must hit the gap exactly, not repeat the whole topic.
Plan the check: how will you understand the gap is closed? Pre-determine a criterion — for example, the learner solves two tasks in a row and explains each step without a hint. Without a criterion it is easy to mistake diligence for understanding.
Return to the tool after practice: insert new answers and see if the focus shifted. Such a cycle "answers — hypothesis — practice — new answers" is the refinement of the first result into ready material.
Practice: three tasks on sequential discounts with different numbers plus one task where percentages actually add up (non-overlapping parts of one sum) to show the difference.
How to Check the Result and What Not to Expect
Check each result item with a question: "On which fragment of my text is this based?" If there is no answer, do not use the point yet. The tool may err, and the interface explicitly states that model responses may contain errors, while queue and availability vary.
Do not expect the service to itself set a score, save history, or compare two runs in a special mode. This is not in it: one result only, it can be copied or downloaded as TXT. Comparison you do yourself, manually.
Also do not count on loading documents, batch processing, export to DOCX or PDF, integration and automatic saving. Everything needed for iteration is just text fields and your own analysis. Volume limits: 8000 characters in the first field and 1000 in the second.
If the result claims a gap that is not in your text, shorten the source material to one indicative answer and repeat the run.
Refining the «Find the Gap» result is not searching for an ideal button, but a sequence of small steps: collect verbatim answers, narrow the goal to a checkable skill, read the output as hypotheses, check them with new tasks, refine the input and run again. Two or three such cycles give a specific list of skills and practice for each. Remember the limitations: the tool analyzes only the inserted text, the result may contain errors, and comparing runs and planning practice remain your responsibility. The percentages example in the article is a educational illustration, not a description of a real learner.
Open tool: Knowledge gap finder ↗Interface translations are machine-assisted and may need refinement.