How good is Chatlayers NLU?

Hi,

I've been wondering about the NLU that Chatlayer uses.

1. For example: if you look at the Dutch training phrases "Hoe schrijf ik me in" and "Hoe kan ik me inschrijven". Both cases deal with "inschrijven", but "inschrijven" is verb that can be split in different parts (as happens in the first sentence "schrijf .... in").
   
   I know Chatlayer can be trained to classify both utterances under the same intent. But... what happens exactly? Does the NLU register the verb in the first sentence as "(schrijf) > schrijven" or is it "familiar" with samengestelde werkwoorden in Dutch and does it register the verb in the first sentence as "(schrijf .... in) > inschrijven"? Or does it do both?
   
   And, if you feed the NLU a set of training phrases in which you alternatively use the split form of the verb and the infinitive, is the NLU able to learn:
   > that not only do all phrases in the set 'belong' to the same intent,
   > but also that the split form and the infinitive are actually the same verb? 
   
   
2. What is the name of the NLU that Chatlayer uses? Are there other platforms that use the same NLU?

Thanks in advance!