This issue is a grey area in regulation (deliberately?) and is open to interpretation.
The operator (Captain) is responsible for safe operation, there is a reminder of this in the OPS section – before landing assessment. There may also be other assumptions in the regulations (including CS25), but the final judgement is that of the Captain before landing.
Providing you can justify the choice of action as being safe then all is well, but if the chosen data is less than the established safety margin based on the dispatch data, then with hindsight after an incident, the decision could be debatable. The debate would be about the equivalent level of safety, and thus that there should not be a gap between dispatch and pre landing.
Many manufactures provide ‘actual’ data which requires considerable knowledge of what the data implies and then requires assessments, additions, or adjustment in correctly recognising contributing factors. If assessed correctly, and appropriate safety margins are added, then the pre landing check using ‘actual’ distances should be close to the dispatch case.
Airbus, for one, now publishes Operational Landing Distances (OLD), more often with a Factor (FOLD) – check what the factor is - which provide more realistic and achievable landing distance for the conditions. There are still provisions such as the reported runway conditions are reasonably accurate (Captain decision again).
In many situations (most?), FOLD should closely agree with the dispatch data, thus the pre landing FOLD data could be deemed ‘safe’.
But then the crew still has to fly the aircraft according to the assumptions in the data.