I got this question in a comment recently.
I’ve been doing some research at chatbots and I agree with what you’re saying here! I’m build a chatbot with Chatfuel, but that is on messenger. Do you know a good bot for websites I can try out?
While I cannot really give a comprehensive answer I can provide a few general guidelines as you try to build out your website chatbot.
Conditional Logic Chatbots
These chatbots are primarily based on conditional logic and don’t have any AI in them. Chatfuel is a good example.
While these are useful in some scenarios they won’t work when you actually have a need to understand what the user is saying. This is quite different from simply providing a menu of options for the user to choose from. If you do plan to use them, here are a couple of services which I have heard good things about – Motion.AI and GoBot. I haven’t tried either of these as I am focused mainly on NLU based chatbots.
My opinion, which is based mostly on helping a few folks with Chatfuel + Dialogflow integration, is that the basic paradigm of a conditional logic based chatbot makes it quite hard to integrate with an NLU based bot. If you are very keen, you can certainly try it out. (Note: not all of them can be integrated into your website with anonymous login).
Adding bot logic to a customer messaging platform
Drift is a popular customer messaging platform which is used primarily for the live chat scenario. You find it on many popular websites, with a chat window that (usually) pops up on the bottom right with a “We’re Powered by Drift” link at the bottom of the chat window. Here is an example from the Drift website itself:
However Drift does support external integrations via their API, which means you can in theory build some logic in Dialogflow and integrate it with your Drift chatbot. I plan to try this out and report back here some day (hopefully soon).
NLU based chatbot
If you want a chatbot with a little bit more AI (NLU-based), you can also integrate a Dialogflow chatbot into your website.
I have written about this quite a bit on my site, and I am in the process of creating a course on this topic which provides all the basic concepts of implementing one on your own site. (I already have a course on this topic, but I found that it is very challenging to keep the sample code in the course up to date).
With an NLU based chatbot, you get the benefits of using Dialogflow to try and understand what your customer is saying. In addition, by using features like context-based fallbacks, you can actually do quite a bit more. (More on this in a future article).
What has been your experience with bots used explicitly for websites? Please leave your comments/feedback below.
- Tips for learning Dialogflow CX
- Dialogflow CX vs ES: First look
- Seven ways to integrate a Dialogflow chatbot into your website
- Dialogflow Zobot: Selection Triggers the next intent
- Dialogflow Architecture
- Dialogflow Python webhook tutorial
- Dialogflow training
- Dialogflow vs Lex vs LUIS vs Watson vs Chatfuel
- Convert your WordPress website into a Dialogflow FAQ chatbot
- When NOT to use follow up intents in DialogFlow