The easiest way to find information in the corpus is to select a (sub-)corpus in the bottom left and then type in the token (word) you are looking for in the top left query box, like "ich" (in double-quotes). This case sensitive type of query will find all tokens* with exactly this pattern: an <i> followed by a <c> followed by an <h>, all of that in lower case and with no other characters in the token. These are examples of tokens that will not be found in spite of them containing the character sequence ich:
Another form that will not be found is Ich, because we query for lower case letters, so upper case will not be found.
If you want to get as a result all tokens containing –ich or all spelling variants, you have to use RegEx.
*To be more precise: the query ich will find all occurrences of this pattern as either a token or a message. Thus, if a message consists of only this pattern, it will be displayed twice, once as a token and once as a message. To avoid this, query for tok="ich" (or msg="ich" if you want messages with only this pattern).