AI Agent for Customer Support That Hands Off When Unsure
An AI agent for customer support has to know when to stop. CustomModel answers from your content, flags answers it cannot ground, and passes the customer to your team with the question attached.
Answers first, people for the rest
The agent resolves the repeat questions in the chat. When the content has no answer, it says so and passes the conversation on, so customers are not stuck in a loop.
- Cited answers
- Grounded flag
- Handoff with the full question
curl \
https://custommodel.ai/api/v1/answer \
-H "Authorization: Bearer $CUSTOMMODEL_KEY" \
-d '{"model": "support", "question": "What time is late checkout?"}'200 OK
Your team stays in charge
You decide what content the agent can use and which tone it takes. Seats let several people manage models, and analytics show where to improve.
- 5 seats on Growth
- Unlimited seats on Enterprise
- Question analytics
Plug it into your tools
On Growth and Enterprise the agent is one HTTP call that returns the answer, citations, and grounded flag.
-
JSON output
-
Works from any backend
-
Same content as the widget
What it costs
Plans start at $24 a month billed yearly ($49 monthly) with 1,000 answers, and Growth at $99 a month billed yearly ($199 monthly) adds the API and five seats. See chatbot pricing for every limit.
Starter: Starter is $24 a month billed yearly ($49 monthly).
Growth: Growth is $99 a month billed yearly ($199 monthly).
Enterprise: Enterprise is $299 a month billed yearly ($599 monthly).
Related pages
Keep exploring: customer service chatbot, RAG as a service, chatbot security.
FAQ
Does it take actions in my systems?
No. It answers questions from your content and hands off to your team.
What if the question is not covered?
The response is flagged as not grounded and the customer can be passed to a person.
Can I call it from my own app?
Yes, on Growth and Enterprise.