An successful organisation is dependent on many factors, innovation, the right product at the right time, it's connection to customers and lots more.
But under it all lies general knowledge and the efficiency in which such accumulated knowledge is shared and used. Human Intellectual Capital for "short".
Let's simplify (a lot), doing a bit of math for the fun of it (Note: No duplication of experience):
Lets say you have ten support people, one has ten years of experience and the other nine have one year of experience each.
If all keep all knowledge to themselves the organisation would have 19 (experience-years * man-years ) as resource in total. (10*1 + 1*1 + 1*1...)
Worse, those 19 years of experience is not easy to use efficiently as nobody would know where to pipe a request or issue. "Who if anybody knows anything about this?"
Now, if the same team shared all, knew all, and used all, you would have 190 (experience-years * man-years) as total resource. (19*1 + 19*1...)
Better, no need to steer the issue and request flow, anybody who is free could start working on an issue.
Today the somewhat structured tasks (support functions, bug fixing etc.) are mostly covered, we're flush with bugtracking software and CRM systems (but one might add, where's the focused learning and efficient knowledge distribution and use in those?).
It's the ad-hoc small and big issues that fills most of the working day. So the effect of capturing what happens there, flipping that into organisation-wide learning and making it easily accessible and useful would have a huge effect. Enormous.
To attain that, the complete process must be captured. Not only the "issue - solution" pairs, the start- and end-points as they are.
Knowledge is about all that relates to an issue; the possible redefinition of the issue, every blind alley explored, every alternative tried and who was involved and did what. Learning involves understanding the process and not so much accumulating digested results. That would be a generally accepted pedagogic principle.
That's why it should be about process "management", not managing the results, aka "knowledge management".
And definitely not about "document management", don't even get me started there.