Most keyowrd databases consist of a small sample of the overall search universe. This means keyword databases tend to skew more toward commercial terms and the core/head industry terms, with slighlty less coverage of the midtail terms. Many rarely searched for longtail terms are not covered due to database size limitations & lack of commercial data around those terms. Plus if those terms were covered, there would be large sampling errors. Google generates over 2 trillion searches per year and claims 15% of their searches are unique. This means they generate searches for over 300 billion unique keywords each year. The good news about limited tail coverage is it means most any keyword we return data on is a keyword with some commercial value to it. And with Google's Rankbrain algorithm, if you rank well on core industry terms then your pages will often tend to rank well for other related tail keywords.
Repeat this exercise for as many topic buckets as you have. And remember, if you're having trouble coming up with relevant search terms, you can always head on over to your employees on the front lines -- like Sales or Services -- and ask them what types of terms their prospects and customers use, or common questions they have. Those are often great starting points for keyword research.
What I really like about them is extensive detail including the keyword difficulty. Ahrefs uses clickstream data to also show how many clicks you will get from the search engine. This is very useful after knowledge graph integration, as many keywords may have huge traffic but they hardly get any clicks from the search engine. Reason being, they get answers directly from the Google search result. Example of one such query is: “birthdate of any celebrity”