ChatCactus — AI support agent
ChatCactus is a live AI support agent for businesses that want customers to get answers at any time of day. Its stated purpose is straightforward: answer customer questions 24/7. You can open the live product and try it before deciding whether to discuss a custom support agent. This is one of the studio’s own products, built and operated by AppliconSoft. The product is presented as a working example of a customer-support agent rather than a promise about results for a particular business.
Product Demonstration Gallery
Product interface and workflow evidence. Add real, dated screenshots from the product team; the following captions describe what each image should demonstrate.
How the AI workflow works
Step-by-step overview of the documented product workflow. AI functionality is distinguished from standard automation where evidence is limited.
Request checks
Validate the message, published agent, allowed website domain, visitor rate limits and workspace quota.
Visitor context
Update page, referrer, language and any previously supplied contact information
Human handoff
to a person when requested or configured; AI remains silent after takeover.
Knowledge retrieval
Embed the question and retrieve up to four relevant document chunks using similarity search.
Prompt assembly
persona, rules, visitor details, cited sources and recent messages.
Model routing
Call the selected OpenAI or Anthropic model, with fallback and cost controls.
Streaming answer
Stream the response and display citations actually referenced.
Follow-up actions
: Escalate when needed or offer lead capture if the knowledge base cannot answer.
Record keeping
Store messages, token usage, model, costs and response times.
Performance and Evaluation
Reported capabilities, measurable indicators and evidence limitations. Vendor claims are not independently verified unless explicitly stated.