01
People know it is AI
Assistant messages are identified as AI. Usherdesk is designed to be useful without pretending that a generated answer came from a person.
Reliable AI at Usherdesk
Useful answers. Visible limits. A person when it matters.
Usherdesk is designed to answer from business-approved content, identify itself as AI, capture only the details configured by the business, and hand the conversation to a person when needed.
Usherdesk AI assistantAI
Human route readyCan I move my viewing to Thursday afternoon?
Answer from approved listing content
Thursday viewings are available from 2pm. Would 2:30pm work for you?Human requested
Full context attachedThe customer asked to confirm with an agent. Conversation and captured viewing time are ready for the team.
Six operating principles
Reliability has to be visible in the product.
These principles guide how Usherdesk is presented and how businesses should configure and supervise it.
02
Answers start with approved content
The business chooses the website pages, documents, and written FAQs available to the assistant. Keeping those sources accurate is part of operating the assistant well.
03
A person can take over
When the assistant cannot help or the customer asks for a person, the conversation can be handed to the team with the available context attached.
04
Customers control the workflow
The business controls source content, lead fields, team access, pipeline stages, channel connections, and where captured leads are delivered.
05
Limits are stated plainly
Generated answers can be incomplete or wrong. Usherdesk should not be the sole decision-maker for emergencies or consequential medical, legal, financial, employment, housing, credit, or insurance decisions.
06
Data use stays bounded
Customer content and enquiry data are processed to provide and secure the service. They are not used to train general-purpose AI models unless the customer separately agrees in writing.
Shared responsibility
Good answers need good operating choices.
Usherdesk provides product controls. The business using the assistant remains responsible for its content, configuration, access, and use of captured information.
Usherdesk provides
The business controls
- Visible AI identity
- Approved-content retrieval
- Configurable lead capture
- Conversation context for hand-off
- Account and team controls
- Accuracy of source content
- Questions and fields collected
- Who can access conversations
- When a person should take over
- Lawful use and customer notices
Before going live
A practical reliability checklist.
- 01Use current, approved source content.
- 02Tell visitors they are speaking with an AI assistant.
- 03Collect only the lead details the business actually needs.
- 04Give customers a clear route to a person.
- 05Test representative questions before launch and after content changes.
- 06Review conversations for recurring gaps, unsafe answers, and outdated sources.
Need to report an AI concern?
Tell us what happened and include the conversation context.