Every fortnight, all the technical people in our section meet for a forum - our chance to bring up any technical issues or things we've learnt that might be of interest to others. This has mixed results, but occasionally it throws up a good idea that the team can pursue.
The general problem with ideas is how easily they can be argued for or against. Of they were implemented, on the other hand, their merit is demonstrable.
Therein lies the difficulty for a scrum team. There's always things that could be potentially of value, but they have to go up against that time being used for something else. The forum becomes an exercise in implicit permission - we put the idea forward for quality control.
If the idea has tacit merit and implicit consent, then it's something that can be put forward as part of the development. Since sprints are about providing business value, the idea should in some way contribute to business value.
One example floating around at the moment is moving to a microservice architecture. We have our reasons for doing this, but it will still take time to set up and deliver. But doing it as part of a particular user story means we can realise that change and assess its merits. It serves the business, as well as allowing for the developers to try out better ways of delivery.