When installed, Propane Chat scrapes / 'reads' all of your website's content to learn everything about your organization. Because of this Propane Chat can already answer most of your website visitor's questions minutes after being installed.
We recognize, however, that having general industry knowledge is a given for many customers and that many products are only valuable in the context of a broader industry. For example, perhaps you rent a vacation home in Portland and want your chatbot to have a general knowledge of the tourist attractions and activities available in Portland.
For these reasons, Propane Chat can be trained not only by your website but also by other websites, or 'external urls', that contain information relevant to your customer's purchasing decision. In the vacation home example one of your external urls might be Portland's tourism website that contains information about things to do there while on vacation. Once the tourism page's url is added to Propane Chat, your chatbot will use the information found on your website and the information found on the tourism website to answer your website visitor's questions. Now, when someone asks about good restaurants in Portland, Propane Chat will have the answer.
To add an external website to Propane Chat's knowledge base, navigate to the Scraping Settings tab in the Propane Chat Dashboard.
Near the bottom of the page you will see 'External URLs':
You can add as many 'Single Page' urls to be scraped as you like. When Propane Chat scrapes these pages it will only scrape that specific url and no other pages on that website. For example, if you entered www.propanechat.com it would scrape our homepage but it would not scrape www.propanechat.com/pricing. This is what you would use if you want information from one specific page of a website to be included in your chatbot's corpus but not any other pages on that website. Another example might be the Portland Wikipedia page. You might want your chatbot to know about Portland but you don't want it to know about every other topic covered by Wikipedia.
You can add up to two 'Entire Site' urls. When Propane Chat scrapes these sites it will scrape the entire website. In the www.propanechat.com example, propane chat would scrape our homepage but then it would also scrape www.propanechat.com/pricing as well as every other page/path on the Propane Chat website. In the Wikipedia example it would scrape the Portland Wikipedia page but then Propane Chat would also proceed to scrape every other page on Wikipedia.com including the pages on quantum mechanics and toe fungus. That's far from ideal. So only use 'Entire Site' urls for websites that uniquely provide information on your industry.
Thanks for using Propane Chat! We hope that you find External URL Training helpful. We look forward to fueling your customer engagement and support one chat at a time.
- The PropaneChat Team
Install PropaneChat for free today: https://app.propanechat.com/auth/signup
Have more questions about PropaneChat or want help setting up your personalized AI Chatbot? Feel free to ask PropaneChat yourself using the chat feature on this website or reach out to us at hello@propanechat.com.