If you collect information from a work-group within your organization, say a 10-person sales team, then you can use the data to say something about that sales team, but the results will not be representative for the entire organization.
If you take a random sample of 10 people throughout your much larger organization, then you can say something about the organization as a whole, but the variability of the scores will be too great to give certainty that you have a good representation of the organization.
If you have a small organization (of say, 10 people in total) then 10 people is all you need to say everything about that organization. However, if one person leaves the organization and is replaced by someone else, the scores are likely to change drastically upon a second run of a survey.