What should the AI agent actually finish for you?
*
Answer questions
Find information
Qualify leads
Book appointments
Process customer requests
Update business systems
Complete a multi-step workflow
What would the agent need access to?
*
Company documents
CRM
Email
Calendar
Business database
ERP or internal software
Other business systems
Select all that apply.
What should the agent be allowed to do without approval?
*
Answer questions only
Prepare work for a person
Update basic information
Send routine messages
Complete low-risk tasks
Complete approved workflows
Nothing without approval
What should happen when the agent is unsure?
*
Ask another question
Search approved information
Stop and wait
Send it to a person
Create a task for review
Use a safe fallback response
It depends on the task
What must always require human approval?
*
Customer communication
Financial actions
Changes to important records
Refunds or cancellations
Sensitive data access
High-impact decisions
Nothing within approved limits
What is the worst mistake this agent could make?
*
Give a wrong answer
Take the wrong action
Contact the wrong person
Expose private information
Lose or change data
Break company policy
Create financial loss
How should we reach you?
First Name
*
Last Name
*
Work Email
*
Role
*
Phone
Do you want to sign NDA
yes
Anything we should know before we speak?
Preferred mode of communication
Email
Phone
SMS