A deductive reasoning solver needs every rule and the exact question being asked. TestSolve reads those details off your screen and returns a worked answer to your phone, keeping the conditions, diagram and answer choices together. You do not have to recreate a scheduling table or arrangement puzzle inside a chat window.
The essential distinction is between an answer that is possible and one that is required. “Must be true” asks what follows in every arrangement allowed by the rules. “Cannot be true” asks what is impossible in any allowed arrangement.
The TestSolve deductive reasoning workflow is built for questions already on a computer screen. You can try the first question free, without an account, and inspect whether the explanation proves the right kind of conclusion.
A deductive solver works from stated facts and conditions to determine which conclusion follows. Unlike an open-ended question about what is likely to happen, the answer is constrained by the information supplied. The tool should not add assumptions simply because they would be realistic.
Some questions use sentences about groups or categories. Others use schedules, positions, rankings or assignments. The surface presentation changes, but the core job is to apply all the conditions consistently.
A worked answer should connect the selected option to those conditions. For a necessary conclusion, it should explain why alternatives cannot overturn it. For an impossible conclusion, it should identify the conflict that prevents it.
Must be true means the statement holds in every valid scenario. Could be true means at least one valid scenario supports it. Showing a single arrangement where a statement works proves possibility, but it does not prove necessity.
Imagine three presentations, with the only rule that Amir speaks before Beth. Cara could speak first, between them or last. An explanation that shows Cara in the middle has demonstrated one possible arrangement, not that Cara must be in the middle.
This is a useful check on an AI response. If the question asks what must happen, look for an argument covering all valid possibilities. A neat example alone may not answer the question.
Cannot be true means there is no valid scenario in which the statement holds. It is stronger than saying the statement does not appear in one example. The solver must identify why it conflicts with the rules in every attempted arrangement.
In the presentation example, Beth speaking before Amir is impossible because it reverses the explicit condition. Cara speaking first remains possible because no rule forbids it. The conclusion depends on the stated constraint, not on the solver's preferred schedule.
The question wording therefore needs to remain visible. A solver that confuses “cannot” with “could” may do competent reasoning and select an answer to the opposite task. Preserving the instruction with the full problem is part of giving the AI usable input.
These phrases require different forms of evidence. The following table is useful when deciding whether a worked explanation has done enough. It applies to the logic of the question, regardless of which tool produced the answer.
| Question asks | What must be established | What is not enough |
|---|---|---|
| Must be true | The statement holds in every valid scenario | One example where it holds |
| Could be true | At least one valid scenario makes it true | An arrangement that violates another condition |
| Cannot be true | No valid scenario permits it | One example where it happens to be false |
| Must be false | The statement is false in every valid scenario | A personal judgement that it seems unlikely |
| Which arrangement is valid? | Every supplied condition is satisfied | Checking only the most obvious conditions |
Under the ordinary consistent-rule interpretation, cannot be true and must be false point to the same logical requirement. Everyday plausibility does not determine either answer. The conditions do.
SHL's published fact sheet for Verify Deductive Reasoning specifies 20 minutes and a maximum of 18 questions. It describes an adaptive multiple-choice assessment, and its example presents five answer options. Those details identify a particular documented format.
They should not be used as a universal description of every SHL deductive product. Interactive versions and combined assessments can differ. Check the name and instructions in your invitation before applying a timing or interface description.
For a solver, five options create five claims to distinguish. Selecting the first plausible statement can be insufficient when the task asks for a necessary conclusion. A useful explanation identifies what makes the chosen option uniquely satisfy the instruction.
ChatGPT can be useful for a short set of clearly written conditions. You can paste the rules and ask for a conclusion with an explanation. Its conversational format also makes it convenient to question an assumption or request a counterexample.
The risk is that the explanation may establish less than it claims. A response can show one valid arrangement, then mistakenly describe a feature of that arrangement as compulsory. It can also overlook a condition that appears earlier in the prompt.
TestSolve offers a different input workflow when the problem is displayed as a grid, table or diagram. The screen reading deductive solver receives the visible problem and sends its working to your phone. The logic still needs to be sound, but you avoid manually rebuilding the display.
A schedule may combine text rules with a visual calendar. Some cells may already be filled, while others are unavailable. If you paste only the written conditions, you can omit restrictions that are encoded in the grid.
The labels can matter as much as the positions. “Before” might refer to an earlier day, an earlier time or an earlier place in a sequence. The complete display helps establish which interpretation applies.
Screen reading preserves those visible relationships as input. It does not supply constraints hidden on another page or guarantee that every blocked cell is recognised. When checking the answer, compare the proposed arrangement with both the written rules and the visual restrictions.
One mistake is reversing a condition. “Every approved request has a reference number” does not establish that every request with a reference number is approved. An AI may accidentally treat a one-way rule as if it worked in both directions.
Another mistake is assuming that a category must contain a member. A statement about all members of a group does not always establish that any exist. Whether existence is given can affect what conclusions follow.
A third is silently adding a uniqueness condition. If the question has not said that each role, day or person can be used only once, that restriction needs justification. A tidy arrangement based on an invented rule can be logically irrelevant to the actual problem.
Read the task word first: must, could or cannot. Then ask whether the explanation proves that requirement. This catches a different class of mistake from checking whether the final arrangement looks reasonable.
For a must-be-true answer, try to imagine a valid scenario where the statement fails. One valid counterexample defeats a claim of necessity. For a cannot-be-true answer, one valid scenario where it holds defeats the claim of impossibility.
You do not have to enumerate everything to spot a clear flaw. Often the decisive check is a single overlooked condition or an unjustified assumption. A solver's working should make that check easier by naming the rules it used.
For a short textual argument, a free assistant may be all you need. For a visual schedule with several constraints, direct screen input can be more convenient. The useful comparison is how well each option receives the whole problem and explains its conclusion.
| Option | Best reason to try it | Main limitation |
|---|---|---|
| ChatGPT | Written rules with room for follow-up questions | Can confuse possibility with necessity |
| Aptitude Solver custom GPTs | A focused starting point for the conversation | Specialist naming is not proof of better reasoning |
| aichatonline.org or yeschat.ai | Free or trial access to aptitude assistance | Actual features and limits need checking |
| TestSolve | Screen questions with phone-based worked answers | Complete visible input and logical checking still matter |
| Heycademy or Arsturn guidance | Understanding formats and tool choices | Guidance is different from solving your exact display |
Microsoft Math Solver was a focused mathematical tool, rather than a general solution for verbal constraints and scheduling logic. Microsoft also announced its retirement for July 2025. It should not be recommended merely because the word “solver” appears in its name.
TestSolve reports 92% on a 50-question benchmark measured on 18 September 2026. That aggregate result does not establish a deductive-only score or a score for SHL. It should not be used to promise a particular assessment outcome.
The product supports Windows, macOS and browser screen sharing. Your first question is free, followed by question packs at $14.99 for 30 or $19.99 for 50. There is no subscription, so the purchasing decision is about whether this workflow is useful for the questions you need answered.
Personality questionnaires and strengths inventories have no universal right answers for the tool to derive. Game-based assessments score how you play, so a worked answer to a still screen does not complete them. Follow the assessment's rules on permitted assistance.
For text rules, ChatGPT or another free assistant can be useful. TestSolve is designed for questions already on a computer screen, including visible grids and tables, with worked answers sent to your phone. Choose by input format and the quality of the explanation.
It means the statement follows in every scenario allowed by the information. A single example where the statement holds is insufficient. If you can construct a valid scenario where it is false, the statement does not meet the must-be-true requirement.
Cannot be true means the rules make the statement impossible. A statement that happens to be false in one arrangement may still be true in another. The solver needs to show a conflict with the conditions, not simply present one unfavourable example.
SHL's fact sheet for the documented multiple-choice format states a maximum of 18 questions in 20 minutes, with five options in its example. Other SHL formats can differ. Use the instructions for the specific module in your invitation.
It can attempt them if the screenshot includes the rules, labels, visible restrictions and answer options. Missing a blocked cell or an extra condition can change the solution. Check the proposed answer against the complete display before accepting it.
Free chat tools can attempt written reasoning questions, and TestSolve offers the first question free without an account. Its paid packs cost $14.99 for 30 questions or $19.99 for 50. There is no subscription to start or cancel.
Keep the rules, diagram and answer choices together so the solver can work from the same question you see. Try TestSolve's screen reading solver and inspect the worked conclusion on your phone. First question free, no account, no subscription.
<!-- Editorial source notes, not part of the article body: SHL format, maximum item count, timing and five-option example: https://service.shl.com/docs/Verify_Deductive%20Reasoning_Fact%20Sheet.pdf Microsoft retirement: https://x.com/MicrosoftMath/status/1931103750753599946 Product facts and benchmark supplied by TestSolve. All logical illustrations are original and do not reproduce publisher questions. No claim that any named employer uses SHL is made. -->
First question free. No account, no subscription.