Don't pursue KM as a goal in itself. Use it insofar as it serves your purposes.
Be pragmatic rather than idealistic. Only use KM approaches that bring maximum benefit for minimum input. Don't spend ages pushing high‐flying ideas and initiatives that will end up costing you more to introduce/implement than you are likely to benefit from them.
If you try something that doesn't work, don't agonise over it too much - that will just lead you into despondency or an endless spiral of reflection on reflection. Instead, note what you can and move on quickly. Try new things and build on what does work well.
Don't get obsessed with building up repositories of knowledge that nobody is ever likely to access in practice. Be realistic - if you want to keep records of learning, good practice etc, fine, but don't get despondent if others don't access them.
Don't rely on knowledge banks as the basis for sharing and dissemination of learning. Let one‐another know who knows what and develop interactive discussion groups, communities of practice etc. where learning is kept fresh, alive and relevant by interested parties.
Develop communities of practice and other interactive fora based on actual questions that actual people are actually grappling with; e.g. "Does anyone know how to..?" This is, in effect, the opposite approach to developing sophisticated knowledge banks and trying to persuade people to access them.
If you do decide to keep written records (e.g. of learning reviews following projects), record key learning points and action recommendations clearly and succinctly. If you want others to use your learning record, use a clear, generic title that others will recognise and identify with and avoid or explain jargon or acronyms.
As a general rule of thumb, keep KM simple, keep it practical and only use methods that have a reasonable chance of securing benefits that outweigh related costs.