See IrisAgent work on a real support desk
Ask the bot, call the voice agent, score a ticket for escalation, tag it, and review how it was handled. Everything here runs on live AI.
AI chatbot and search
Ask it anything about IrisAgent. It answers only from irisagent.com and docs.irisagent.com, cites the page it used, and hands off to a human when it can't answer confidently.
Chat with the bot
Open the chat bubble in the bottom right corner of this page. Try one of these:
AI search
The same answers, embedded in a help center search bar.
Voice AI agent
Call the agent from your browser or your phone. It answers questions about IrisAgent from the same knowledge base as the chatbot and transfers to a person when it should.
Or dial +1 (949) 942-9274
Live transcript
The conversation appears here once the call starts. Try asking what IrisAgent does, how it connects to Zendesk, or to speak with a person.
Escalation score
Paste a ticket thread, or load a sample, and see how likely it is to escalate and why. It uses the same signals, sentiment model, and weights IrisAgent runs on live tickets. Nothing is saved.
AI tagging
IrisAgent reads each new ticket and tags it from your own tagging procedures, so it reaches the right team with the right priority. It uses the same ticket as the other tabs. Nothing is saved.
Tagging reads the subject and what the customer wrote.
The ticket is tagged against these plain-English procedures, one tag per field. In IrisAgent, your team writes its own or generates them from past tickets.
AI QA review
IrisAgent grades every support conversation against your QA rules, not a small sample, and explains each pass or fail so coaching is specific. This uses the same ticket as the escalation tab. Nothing is saved.
Internal notes aren't graded, since the customer never saw them.
Each conversation is graded against these 7 rules. The score weights each rule by its severity. In IrisAgent, your team writes its own rules in plain English.
Want to see this on your own tickets?
Connect your helpdesk and we'll run the same models on your conversations.
Book a call