Hi! ​@EFree Please send me a DM with your account details so we can review this with our internal team. We want to ensure the AI is processing your uploaded information correctly and will look into this further.

Hello! I started using it this week. Any luck on getting it to correctly respond to the information you uploaded to the knowledge hub? What I'm running into so far is that it appears to give priority to any info you have in the FAQs over the knowledge hub. 

Same here! Want to get my receptionist in production. What would help all of us:

1. A sandbox to test what we’re building, not at the individual stage level, but in basically a production level sandbox. 
    
2. The 500 character “prompt” (ie, company description) is way too short. It also may be better used as a traditional system prompt rather than an faq generating action (“You are a helpful knowledgeable and kind assistant for company XYZ… Always use {variables}. Never say {variable}. Do not insert stock phrases like “would you like me to transfer you to X”. Trying to build a capable assistant not a switchboard. 
    
3. Please provide more information on the knowledge base usage, and how the AI references it. 
   we can create (some of us have this handy) our own faq; can the knowledge base contain instructions on bot behavior…tone/temperature etc? (More than the options for the default voice and style)?
4. Will growing knowledge bases create significant latency in the AI response?
5. are .md or .xml files usable in this feature? Those seem to be the standard with many agentic systems (as well as the current RC IVR config file format). 
    
6. Are key value pair (FAQ for example ) lists better in a csv file or a text/pdf file?
7. An off ramp for non English callers. There has to be a way for the agent to recognize non-English, and have its own path. 
    
8. What benefits/limitations exist if the knowledge base contained a list of URL’s (that basically link to another knowledge base that relates to this one? Can the current version of AiR perform a real time web search like openAI, Gemini and Perplexity?
9. it would be helpful to know the underlying infrastructure of the RC AI build. 

I can go on.. but if anyone out there has any answers advice or creative workarounds, would love to hear them!

Can fully relate with you on all points mentioned above. 3 and 4 are key for what I am working on right now. I have a feeling that having both FAQ/Knowledge base can sometimes overlap or interfere with each other, so would be very beneficial if RC could provide a more in-depth explanation of how both work. 
You also mention non-English callers, which I haven’t even contemplated yet, but definitely an important part of our Contact Center as approx 30% of the volume is from non-english callers.  
 

I will keep an eye on this thread to see if it grows as more RC users start adopting the AI receptionist and even better if we can have an RC representative pitch in to solve some of the questions/unknowns we have. 

I will start testing next week and will provide any insights that I may find useful.

Thanks! 

​@jlondono I just signed up.  Has the knowledge base improved?  I’ve not applied the agent to any schedule yet, but I’m testing and the only response I get back is related to the business information I fed it when setting up AIR.  I gave it our URL but it doesn’t appear to be able to scan through our various pages.  

Hi Mike, I stopped using it because it wasn’t meeting our needs, and from what I’ve seen, it hasn’t improved much since I last tried it a few weeks ago. Hopefully, they continue enhancing it — it has the potential to be a great tool, but I don’t think it’s quite there yet.

Hi ​@EFree ​@Crystal_ I have found struggles in the knowledge hub as well. It was supposed to take Word docs but it didn’t “see” them at all. I did notice that if I changed them to PDF’s and re-upload it recognized them. HOWEVER - that has not been consistent. It worked with super-basic data like the names and titles of my executives- but not instructions on how I wanted it to handle questions and responses. I will test it and it will not respond properly- but then I will say, “Why did you respond that way” and it will then reference the correct info. Or I will say “please reference your knowledge base” and then it finds the right info. But upon the next test chat- it has forgotten it all again.  I am finding the AI to be very inconsistent in its answers.  I Need Instructions on how to prioritize training the AI. I need to be able to provide better instructions on call routing to groups and teams vs individuals, regardless of what their title is listed in RC. It needs to be smarter with the semantics of what callers ask for. (i.e. various insurance terms).

Hello ​@kboehmer,

I noticed you opened a case about this a few weeks ago, but it looks like it was closed due to no response. If you'd like, I can help resubmit a ticket on your behalf so you can get the support you need.

Just let me know and I’m happy to help!