# AI chats and agents: live demo | Kemeny Studio

> Six live AI demos in one page: sales agent, support agent, document review, internal search, call QA, and output evaluation. Each one shows its tool calls so you can see what it actually does.

An interactive explainer of how AI chats and AI agents differ, and which one a given problem needs. A mode selector switches the agent you are talking to, and a side panel streams every tool call it makes: what it retrieved, what it extracted, what it flagged, and why.

The six modes:

- **Sales agent** (Aurora): the live agent that sells for Kemeny Studio. Reads the team's real Google Calendar, books meetings, remembers you between visits, and can continue on WhatsApp. Same pattern as [Comercial AI](https://kemenystudio.com/comercial-ai).
- **Support agent**: searches a help center before answering, opens a ticket or escalates instead of guessing. See [Knowledge Management](https://kemenystudio.com/solutions/knowledge-management).
- **Document review**: reads a contract, extracts commercial terms with confidence scores, flags risky clauses. See [Document Intelligence](https://kemenystudio.com/solutions/document-intelligence).
- **Internal search**: answers over company documents and cites the exact source section. Takes no actions. See [Knowledge Management](https://kemenystudio.com/solutions/knowledge-management).
- **Call QA and compliance**: scores a call transcript against a rubric and compliance rules, quoting evidence. See [Voice Analytics](https://kemenystudio.com/solutions/voice-analytics).
- **AI output quality**: an LLM judge that grades another AI's answer for faithfulness, completeness and tone. See [Model Evaluation](https://kemenystudio.com/solutions/model-evaluation).

Only the sales mode touches a real system. The other five run on fictional demo data and store nothing the visitor types. Deep-link a specific mode with `?mode=sales|support|documents|knowledge|voice|evaluation`.

The distinction the page teaches: a chat answers questions, an agent takes actions with consequences. Most buyers ask for an agent when a cited answer is what they need.

---
Canonical page: https://kemenystudio.com/chat
Site overview for LLMs: https://kemenystudio.com/llms.txt · Full content: https://kemenystudio.com/llms-full.txt
