Train AI agents on your own data so they actually know your business. Plug in your docs, FAQs, and knowledge base, and they’ll handle customer questions with fast, helpful answers 24/7 - all while saving you time and support costs.
Train AI agents on your own data so they actually know your business. Plug in your docs, FAQs, and knowledge base, and they’ll handle customer questions with fast, helpful answers 24/7 - all while saving you time and support costs.
Yes, you can leverage multiple knowledge bases for training! Relevance is prioritized using semantic-based ranking, ensuring the most contextually appropriate information takes precedence.
This is really clever and such a breath of fresh air compared to those generic bots that just go in circles. Does the AI get better over time by learning from real customer interactions or does it stay the same once it’s trained?
@charlotte_richardson1 Thanks for the kind words! Currently, you can retrain the AI using our FAQ feature, which we prioritize for quick updates. We're actively working on enabling retraining from real customer interactions, and you can expect that feature soon. Stay tuned!
we enforce a strict system prompt and allow custom prompts for tailored control. Our proprietary evaluation techniques further ensure accuracy and alignment. We also have fallback logic to handle ambiguous inputs gracefully, maintaining consistent and on-brand outputs.
@harper_young Thanks for the enthusiasm! Currently, our AI operates as a standalone support layer, but integration is on our roadmap and coming soon. Check out our integration page for the latest updates!
@noah_ortiz Our AI relies on your data preparation to ensure accuracy. If conflicting or outdated information arises, you can retrain the with updated data to resolve inconsistencies and keep responses current.
Replies
Train AI agents on your own data so they actually know your business. Plug in your docs, FAQs, and knowledge base, and they’ll handle customer questions with fast, helpful answers 24/7 - all while saving you time and support costs.
Does the system have version control for uploaded documents so the AI can recognize when something has been updated or is no longer in use?
@alexander_green6 Thanks for the feedback! We're working on implementing version control. Expect it soon!
Can you use multiple knowledge bases for training?
If so, how do you prioritize relevance among them?
@abigail_phillips2
Yes, you can leverage multiple knowledge bases for training! Relevance is prioritized using semantic-based ranking, ensuring the most contextually appropriate information takes precedence.
@iyared Do you provide multilingual support right off the bat or would that need separate training data for each language?
@benjamin_anderson3
Yep, absolutely! Multilingual support is baked right in from the start—no extra training data needed for each language. 😎
I tried it out with a small FAQ, and it performed better than I expected!
@brooklyn_campbell Thanks for giving it a try, please let do let us know if you have any comments!
This is really clever and such a breath of fresh air compared to those generic bots that just go in circles. Does the AI get better over time by learning from real customer interactions or does it stay the same once it’s trained?
@charlotte_richardson1 Thanks for the kind words! Currently, you can retrain the AI using our FAQ feature, which we prioritize for quick updates. We're actively working on enabling retraining from real customer interactions, and you can expect that feature soon. Stay tuned!
How do you avoid hallucinations or off-brand responses especially when the uploaded data is sparse or unclear?
Do you have any fallback logic in place?
@daniel_reed3
we enforce a strict system prompt and allow custom prompts for tailored control. Our proprietary evaluation techniques further ensure accuracy and alignment. We also have fallback logic to handle ambiguous inputs gracefully, maintaining consistent and on-brand outputs.
Is the onboarding process non technical?
Can support teams train it without needing developer assistance?
@gabriel_phillips Yes.
Amazing concept. Can this work with CRM tools like Zendesk or Intercom or is it meant to be a standalone support layer?
@harper_young Thanks for the enthusiasm! Currently, our AI operates as a standalone support layer, but integration is on our roadmap and coming soon. Check out our integration page for the latest updates!
Training AI with your data is perfect for accuracy. How do you deal with conflicting or outdated information in a company’s document set?
@noah_ortiz Our AI relies on your data preparation to ensure accuracy. If conflicting or outdated information arises, you can retrain the with updated data to resolve inconsistencies and keep responses current.
What formats do you accept for training PDFs, markdown, URLs?
@victoria_anderson5 Thanks for the feedback! We support training with PDFs, TXT, DOC, and URLs, including web crawling for URL-based content.