There are a couple of things you could try.
You can use summary searches, and then in your dashboards search the summarized data. You'll need to include any fields you might want to search by in your summary search, otherwise that field would be lost.
http://docs.splunk.com/Documentation/Splunk/6.3.3/Knowledge/Usesummaryindexing
Accelerated data models will in general make dashboard and reports against the data model faster.
http://docs.splunk.com/Documentation/Splunk/6.3.3/Knowledge/Acceleratedatamodels
You could invoke your saved searches on a schedule, and then use the loadjob command to load the results and pipe them to another search/transforming command using your tokens.
http://docs.splunk.com/Documentation/Splunk/6.3.3/SearchReference/Loadjob
You could try to use accelerated searches as your primary search and then use the tokens in post-process searches in your panels. There's a section on using post-process searches in the link below. This would probably only work if a) your base accelerated search returned transformed data (ie, a stats table), b) you were selecting out a subset of rows using a token, and c) you needed multiple panels referencing the same same base search.
http://docs.splunk.com/Documentation/Splunk/6.3.3/Viz/Savedsearches
Which option you choose depends on what your searches look like, how much work you want Splunk to perform in the background, and how much control you want over the acceleration.
... View more