Numerical reasoning (Aon scales numerical)
37 statements · 12 minutes (roughly 19 seconds per question)
What it tests. Information-retrieval speed, data synthesis and rapid mental arithmetic
Worked example. A statement claims APAC revenue declined every year 2023-2025 in USD; you open the Regional Sales tab, check whether figures are local currency, cross-reference the FX tab if needed, and judge the trajectory.
Common traps. Trying to prove a Cannot Say mathematically, misreading labels (millions vs thousands, fiscal vs calendar year), and reading all tabs before answering.
How to handle it. Read the prompt first, identify the key metrics, click straight to the relevant tab. If a statement needs a three-step calculation, guess fast or select Cannot Say and move on. Speed beats perfection.
Verbal reasoning (Aon scales verbal)
49 statements · 12 minutes
What it tests. Selective reading, semantic differentiation and deductive logic
Worked example. A compliance tab says AML training is due within 30 days except for transfers who completed equivalent training within 12 months; the prompt asks whether a London-to-NY transfer must complete it within 30 days.
Common traps. Importing outside knowledge, and missing restrictive qualifiers like "only," "never," "except," or a buried negative.
How to handle it. Treat the text as a strict contract. If the prompt asserts more than the text supports (intent vs exact headcount), the answer is Cannot Say. Scan for prompt keywords rather than reading sequentially.
Deductive logical reasoning (Aon SwitchChallenge)
5-6 minutes
What it tests. Fluid intelligence, deductive reasoning and operational working memory
Worked example. If input [Circle, Star, Square, Triangle] becomes output [Triangle, Circle, Star, Square], the operator is 4-1-2-3.
Common traps. Drawing the tracks on paper (too slow) and fixating on a hard multi-tier level instead of guessing to clear it.
How to handle it. Track one distinctive symbol (e.g. the star) from input to output; finding where it lands usually eliminates three of four options. For a double-layer operator, work backward from the output.
Inductive logical reasoning (Aon GridChallenge)
6-9 minutes
What it tests. Spatial short-term memory and executive cognitive control
Worked example. A 5x5 grid shows three blue dots in an inverted L; the grid disappears, you judge whether a split diamond is symmetrical, then click the exact three coordinates.
Common traps. Over-investing in the distraction task so the dot memory decays, and missing a 90-degree grid rotation.
How to handle it. Translate the dots into a verbal shorthand (an imaginary 1-9 keypad) and hold that mantra while clicking through the symmetry check; treat the distraction as a binary toggle.
Situational Judgement Test (Aon chat-based SJT)
15-20 minutes
What it tests. Professional judgement, risk management, prioritization and culture fit
Worked example. An associate flags a calculation error in a pitchbook the VP already signed off, with the client meeting in two hours, and asks what to do.
Common traps. The lone-ranger (solving a big issue alone without escalating), the over-escalator (running to an MD or HR over minor peer issues), and the perfectionist (rebuilding a model from scratch mid-crunch).
How to handle it. Prioritize client transparency and accuracy over saving face; present the solution alongside the problem, check your work, consult your associate and keep stakeholders informed.