Receptionist
Main number or after-hours route
Answers approved questions, captures intent, and transfers.
AI agents
AI agents that answer, qualify, book, and hand off
Put an AI agent on a number, an IVR option, a time condition, or a queue. It answers from your knowledge and transfers to your team with the conversation kept in the same workspace.
Direct answer
What can Callfora AI agents do on a live call?
They answer inbound calls from knowledge you approve, qualify and route, book a configured appointment, invoke a tenant-owned HTTPS action you define, and hand off to a person. They do not run outbound campaigns or act as a general autonomous employee. Call Centers or Enterprise, the product module, usage limits, and the speech gateway still apply.
The job
Role, task, and inspectable outcome.
Each tab is a boundary you can verify in the product before a live number is attached.
A generic bot cannot be trusted on the main business number.
Start from a role, then keep a human fallback.
Pick a receptionist, sales, support, or custom job. Templates are starters, not a general skills framework.
- Role starter to launch
- Instructions you write
- Fallback extension on every agent
Available on Call Centers and Enterprise. AI usage and deployment configuration apply.
AI agent setup
The caller needs a booking or a system action, not another menu.
Complete only the task you configured.
Book a calendar slot when locations and a booking user are set, or invoke a tenant-owned HTTPS action you approve.
- Configured calendar booking
- Tenant-owned HTTPS actions
- No native CRM connector implied
Available on Call Centers and Enterprise. AI usage and deployment configuration apply.
Team calendar booking
Nobody can see what the agent said, booked, or passed to a person.
Inspect the session and the usage.
Review transcripts, action results, handoffs, and minutes used against the limits your administrator set.
- Sessions and transcripts
- Handoffs and action results
- Usage limits in the same workspace
Available on Call Centers and Enterprise. AI usage and deployment configuration apply.
Desktop capture
Desktop · human handoff with conversation context
Insights and usage
Unclear job · 1 of 3
The operating model
A visible path from the first hello to a measured outcome.
AI is not a separate bot project. It is a configured participant in the same call flow as your team, with a defined job, explicit boundaries, and a person ready to take over.
- 01
Role
Choose the job and the conversations it should handle.
- 02
Knowledge
Approve the facts it may use in an answer.
- 03
Placement
Put it on a number, menu option, schedule, or queue.
- 04
Task
Configure booking or a bounded HTTPS action.
- 05
Handoff
Name the person or queue that takes over.
- 06
Outcome
Review the conversation, task result, and usage.
Role-based setup
Start from a job, then make it yours
Pick a role starter, write instructions, and keep a human fallback. Custom roles remain supported.
Available on Call Centers and Enterprise. Availability depends on plan, modules, permissions, and deployment configuration.
Knowledge you control
Agents answer from the knowledge you approve
Keep shared or per-agent facts, then test in chat before a live call. AI responses can be wrong; review what you publish.
Placement
Put the agent on the same routing surface as people
Direct number, IVR option, time condition, or queue seat. Routing stays in one admin portal.
Queue participation requires the Call Centers plan and the relevant contact-center configuration.
Individual placement
Give the agent its own internal number and route a public number or call-flow branch to it.
Queue placement
Add it to an eligible queue so AI and people can answer within the same service operation.
Configured tasks
Book appointments or invoke an approved HTTPS action
Calendar booking needs a calendar owner, locations, and available slots. Custom actions are tenant-owned HTTPS definitions — not named CRM connectors.
Governance
See the conversation and the outcome
Review sessions, transcripts, action results, handoffs, and AI-versus-human analytics. Administrators set usage limits.
Learn more →Human and AI
Measure AI agents the way you measure your team
Queue share, completed tasks, failed actions, and minutes used — labeled with the configuration they depend on.
Built around real work
Four roles. Clear entry points. One human backstop.
Each blueprint narrows what the agent should say and do. Your team chooses the route in, the task boundary, and the destination when judgment belongs with a person.
Sales qualifier
Sales number, menu option, or queue
Collects configured details and routes an inbound prospect.
Support triage
Support number or queue
Gathers context, answers from knowledge, and hands off.
Appointment scheduler
Booking line or routed call
Checks the configured calendar and books a confirmed slot.
Usage limits
Administrators set the ceiling before the agent takes live minutes.
Usage limits live with the workspace. Analytics show minutes used against the plan, completed tasks, failed actions, and handoffs to a person. Public pages being visible does not turn AI on for every workspace.
Desktop capture
Desktop · AI usage limits
Desktop capture
Desktop · human handoff with conversation context
Reserved captures for usage limits and the human-handoff session. IMAGE BRIEF: AI-01 remains the editorial companion when supplied.
What makes it different
AI works inside the operation your people already run.
The same numbers, menus, schedules, queues, handoff destinations, conversations, and analytics support both kinds of agent. A caller does not have to enter a separate AI channel, and a supervisor does not have to reconstruct the result in another reporting system.
AI stays bounded
It answers from approved knowledge and performs only the task you configured.
People keep judgment
A fallback extension or queue remains part of every launch review.
Designed boundary
Useful because the job is specific.
Callfora agents handle inbound voice work within the instructions, knowledge, routes, and actions you approve. They are not general-purpose autonomous workers.
Configured to
- Answer and qualify inbound calls
- Use approved business knowledge
- Book against a configured calendar
- Transfer to a person or queue
Not presented as
- An outbound campaign system
- An unbounded autonomous employee
- A replacement for human judgment
- A system that learns policy from live calls
Six proofs
What you can verify in the product
Proof
Role-based setup
Start from receptionist, sales, support, marketing, automotive, or a custom job. Templates are starters, not a general skills framework.
Proof
Knowledge and test chat
Agents answer from knowledge you write. Test in chat before the agent takes a live call.
Proof
Placement on a number or flow
Attach the agent to a DID, internal number, IVR option, or time condition — the same routing surface as people.
Proof
Individual or queue
Run one agent on a line, or seat it in a queue next to human agents with concurrency you set.
Proof
Configured task completion
Book a calendar appointment when locations and a booking user are set, or invoke a tenant-owned HTTPS action you approve.
Proof
Conversation and analytics
Inspect sessions, transcripts, handoffs, action results, and AI-versus-human reporting in the same workspace.
Appointment booking
- 01
Caller asks for an appointment
- 02
Agent collects name, contact details, time, and location
- 03
Agent reads them back and books the team calendar
Operational boundary: Booking needs a calendar owner, configured locations, and available slots. Incomplete details are never booked.
Approved HTTPS action
- 01
Administrator defines a tenant-owned HTTPS action
- 02
Action is tested, then bound to an agent
- 03
Live invoke needs the matching gateway contract
Operational boundary: No native CRM connectors. Destructive methods always require confirmation. Live in-call invoke is unverified until the external gateway matches this contract.
AI responses can be wrong. Review the instructions and knowledge you provide, and tell callers when they are speaking with an automated assistant where the law requires it. See the AI agents policy and calling compliance notes. AI agents are available on the Call Centers and Enterprise plans. Availability depends on plan, modules, permissions, and deployment configuration.
Questions
Questions about AI agents
What can Callfora AI agents do?
They can answer inbound calls, qualify and route, use knowledge you approve, book a configured appointment, invoke a tenant-owned HTTPS action you define, and hand off to a person. They do not run outbound campaigns or act as a general autonomous employee.
Do AI agents work on the same phone system as my team?
Yes. You place an agent on a number, IVR option, time condition, or queue. Conversations and analytics stay in the same workspace. Live voice also needs the speech gateway deployed for that workspace.
Which plan includes AI agents?
Call Centers and Enterprise. The product module, user permissions, usage limits, and deployment configuration still apply. Public pages being visible does not turn AI on for every workspace.
How are usage limits enforced?
Administrators set usage limits for the workspace. Analytics show minutes used against the plan, completed tasks, failed actions, and handoffs to a person.
AI agents rollout
Put an AI agent where your calls need one.
Let’s find the right way to connect your business.