It is well-known that in auto racing, having another car BEHIND you gives an aerodynamic benefit. The benefit to the lead vehicle is much smaller than the benefit to the trailing vehicle, but having another vehicle behind you a) reduces turbulence and b) avoids a low-pressure area which would have the effect of sucking you back.
In cycling, a quick Google search doesn't give a consensus estimate for the size of the effect, or even whether it's large enough to be measurable at all. That said, the people who argue the effect does exist tend to suggest a power savings of 2-3% (so a time savings of maybe 1% or a bit less). This "unplanned experiment", by a well-regarded sports scientist, gave a time savings of 0.68% (1.5 seconds in a 3:42 effort):
Jess Hull was right behind Faith Kipyegon in the Paris 1500m, and stayed there for almost the entire race. Kipyegon broke her own world record by 0.03% (0.07 seconds). It's a tiny difference; could Hull's presence have been the difference between Kipyegon resetting the world record and a very near-miss?
(Of course, Hull got a far, far larger benefit from her drafting than Kipyegon possibly could have. And there are many other factors that make a larger difference than 0.03% in any race, such as psychological effects and the drafting--or lack thereof--of the hired pacemakers. My focus on Hull's effect on Kipyegon is mostly for fun).
