The subject has been debated in other posts before, so this will be a very short answer. For graphs and pictures, a search in this forum will do.
O2 sensor emits a voltage that reflects the AFR. The voltage is by no means linear. Far from it, and the values above and below 14,7 (the ideal number with respect to emission) are also influenced by the sensor temperature...
Even so, for all practical purpose, it is possible to use the narrow band sensor to conduct a controlled enrichment based on feedback from O2 sensor. Even if it will no be dead accurate when comes to exact improved AFR, it works well enough to work.
The AF-XIED strive to get a steady voltage reading from the O2 sensor, and it does so by controlling the signal to the ECU, and replace the O2 signal by a generated signal that bumps lean pulses to the ECU O2 sensor input in order for the ECU to go richer than what the mixture would have been if the ECU received the direct signal from the O2 sensor.
AF-XIED relates to the voltage from the O2 sensor, and the adjustment selector on the unit sets different voltage levels. This means that whatever the setting is selected, this voltage (and therefor the AFR) will be maintained over the entire rpm range. Hence, any extra lean rpm range will be compensated for.
Even if the ECU use O2 sensor feedback when in closed loop operation (driving with steady power), the corrections for a richer fueling will be stored in a separat table that is considered at all riding conditions. The target values for this table is such that the realtime correction from the O2 sensor should be 0, hence it takes a while for the ECU to fully rebuild the table after the AF-XIED has been installed.
Also, keep in mind that when the throttle is fully opened up, even the OEM fueling is richer, so do not expect any improvement at wide open throttle. The only noticeable effect when opening the throttle may be that as the adaptive values have been richened up, the immediate respons my improve, making the engine feel a bit more responsive.