Once the FAQs are extracted from the URL, build a Knowledge Graph by adding terms in the left pane. You could then drag and drop the FAQs which are extracted from the source (URL) to the terms .
When the user provides an utterance to the chatbot, the KG engine would look out for the number of terms present in the utterance and quality the paths depending on the thresholds defined under Natural Language section. (Would be 50% by default).
Once a path is qualified, all the FAQs at the path would be indexed to pick the right FAQ based on the user utterance and provide the answer.
For each of the FAQ, the developer could configure alternate questions as well for the primary question. Maybe this should answer your query.
You could configure the synonyms at term, tag (added at the FAQs directly). Alternatively you could define Global synonyms instead of path level synonyms which would apply to all terms in the KG at different paths with the same name.
Please refer to below links for more details on Knowledge collection: