26 karma · joined February 13, 2019
It's important to reduce hallucinations, and with a good prompt it's possible to reduce them.
You basically tell the LLM to avoid answering if the context is not enough.
So, most of the time the Chatbot will correctly reply, and that's already a big advantage.
With Userdesk you can supervise the conversations, and the Chat widget automatically asks for the email of the user, so you can always provide support via the live chat or email later on.
At the moment the main diversification is about using Notion as the data for training (along with sitemap.xml and PDFs).
I'll probably get a better idea by talking to the users.
Competition is high, but I think there's a space for every product, and if those already present are successful, it means that the product is already validated.
I'm a SaaS founder, and in the live chat, I often find myself replying to the same questions over and over.
The AI Chatbot would be able to replace a human for most of the users' questions, and you'll be able to step into the conversation if needed (also later via email).
Userdesk currently supports training a Chatbot using Notion, the sitemap.xml of a website, and PDF files.
I'd really love to get your feedback
In this article I describe the process I followed to validate it, and accidentally collect $3k.
I'll check it out.
And it seems to be the best way to approach it.
That's why I tested the feature myself before building the product.
The results had been very good, so I started developing. A lot of users are very happy with that and had been able to kickstart many meaningful conversations
You'll be surprise to discover that actually many people reply back, and a few unfollows happened due to that message in the last 6 months.
It's suited to everyone who wants to kickstart conversations, the welcome message it's just an icebreaker, as you jump into the conversation after the reply.
It's for indie hackers, product builders, SaaS founders, agencies, and so on.
How do you measure engagement?
I totally understand your first reaction "this is tool for marketers willing to use automation tools to bother people" but actually it's not like that.
The tool internally fosters a collaborative and helpful behavior, and clearly asks to avoid spammy content.
The users of Hivoe are willing to kickstart conversations! And that's the way I've been using it since the beginning.
If you behave differently, like in a spammy markerting only oriented way, the users will probably be likely to mark you as a spammer on Twitter. End of the account.
I hope you got my point. Nice to share our opinions!