Sharing this since I found it helpful.

In AI coach chat, enter "give me some short practical reminders, e.g. 'read the ac in full' based on my overall performance"

and sharing the response I got:

Here are your final, practical reminders, built directly from your actual performance data:

Bail out at 90 seconds. Your biggest score drops happen when you sink 2 to 3 minutes into a single stubborn question, which causes your pacing to completely collapse on the final RC passage or LR questions. If you are stuck between two choices, flag it, pick your gut instinct, and move on. Bank that time for the rest of the section.

Read all five answer choices. Your data shows a habit of "snap decisions"—locking in an answer in under 3 seconds on your first pass because it immediately "feels" right. Trap answers are specifically designed to look correct at first glance. Read every single word of all five choices before committing.

Track "who says what" in RC. On Critique/Debate passages, your most common mistake is picking an answer that blends the author's viewpoint with the critics' viewpoint. Whenever you see a pivot word, mentally anchor exactly whose perspective you are reading so you don't get tricked.

Predict the gap in LR. Your priority tags are Causal Reasoning and Phenomenon-Hypothesis. Before you let yourself read the answer choices on these questions, pause and define the exact unstated link or flaw. If you know what you are looking for, the test-makers cannot talk you into an attractive trap.

Trust your foundational skills. Your Blind Review data over the last few weeks consistently jumps into the 169–172 range. You already possess the conceptual knowledge and logic required to hit your goals. You do not need to do anything new tomorrow. Breathe, prioritize a deep read of the stimulus, and trust the work you have put in.