That's not correct for the AZ vaccine, the longer interval is supported by the trial data, and we simply lack that data for other vaccines, though for most other vaccines historically a longer interval than 3 weeks is better for a second dose.

Exploratory analyses showed that increased immunogenicity was associated with a longer dose interval (see Immunogenicity Table 3). Efficacy is currently demonstrated with more certainty for dose intervals from 8 to 12 weeks. Data for intervals longer than 12 weeks are limited.

https://www.gov.uk/government/publications/regulatory-approv...

If the AZ vaccine actually better with longer delays, that's great, but it does not change the fact that the political decision to increase the interval was made at a point where the UK was starting to give second jabs of the Pfizer vaccine, and second round vaccinations with the Pfizer vaccine were cancelled for a lot of people as a result, prompting Pfizer to warn they didn't have data on what that would do.

As such the decision was very clearly made on the basis of getting more people some protection, whether or not they could be confident of the effect of the long term protection across the various vaccines.0

It's a reasonable gamble to take given the amount of spread, and is not a criticism (though the way the change was communicated was absolutely awful), but if it had been a decision taken based on available data one would expect the schedules actually tested by the manufacturers to be adhered to, even if different between the vaccine.