Step 1: Recall the fixed properties of a normal distribution, which are independent of the actual mean and SD values.
Step 2: The proportion of data within $k$ standard deviations follows the 68-95-99.7 rule: $\mu \pm 1\sigma \approx 68\%$, $\mu \pm 2\sigma \approx 95\%$, and $\mu \pm 3\sigma \approx 99.7\%$.
Step 3: Matching the asked limit of one standard deviation directly gives roughly two-thirds of all observations. \[\boxed{68\%}\]