Managed AI agent role
Inquiry & Booking Agent for one measurable queue.
Handle repeat availability, service, pricing, qualification, and scheduling questions without handing sensitive exceptions to an unsupervised bot. Shoris scopes the approved knowledge, connected systems, booking action, and human handoff before launch.
Scope before software
What a credible first deployment must define.
Workload the agent owns
The role is limited to the repeatable path from a new enquiry to an approved next step.
- Answer approved service, availability, process, and price-range questions
- Collect missing qualification and scheduling details
- Offer available slots and confirm the booked next step
- Write the conversation outcome to the connected record
Systems it may need
Connections depend on the workflow you already use. Access remains limited to the fields and actions required for the role.
- Website forms, WhatsApp Business workflow, email, or support inbox
- CRM or lead sheet used as the source of record
- Calendar, booking tool, or appointment inventory
- Approved service catalogue, policies, and response examples
Guardrails and escalation
The agent should never improvise commercial commitments or continue a sensitive conversation beyond its authority.
- Uncertain answers, complaints, refunds, and unusual requests route to a named owner
- Discounts, custom quotations, and contractual claims remain approval-gated
- Low-confidence identity or booking details require confirmation
- Logs preserve the answer, source, action, and escalation outcome
KPIs and starting scope
Begin with one enquiry source and one booking action. Compare performance with the existing operating baseline before expanding.
- Median first-response time and percentage answered within target
- Complete enquiries converted to confirmed next steps
- Human escalation and correction rates
- Start scope excludes unrestricted sales negotiation, payments, and unsupported channels
Fit check
Know when to deploy—and when to wait.
The right first workload is repetitive, measurable, digitally accessible, and has a named human exception owner.
Good first scope
- The same questions arrive every week
- Availability and booking data is digitally accessible
- An owner can define approved answers and exceptions
- A response-time and booking baseline exists or can be measured
Not ready yet
- Most enquiries require bespoke commercial judgment
- Availability lives only in staff memory
- No person owns escalations
- The expected volume is too low to justify the integration effort
Buyer objections
Questions to resolve before signing.
Will customers know when they are speaking with an agent?
Disclosure should match the channel and use case. The operating design should not misrepresent the agent as a person, and a human route must remain visible for sensitive or unresolved cases.
Can it quote every service automatically?
Only approved, structured price information should be automated. Custom quotes, discounts, and exceptions remain with the commercial owner unless a separate controlled approval workflow is designed.
What if our calendar or CRM data is messy?
The first scope should use only reliable fields. Data cleanup, record deduplication, or system migration is separate work and should be priced before the agent depends on it.
Does the deployment include WhatsApp charges?
No. Messaging provider, model/API, telephony, infrastructure, data cleanup, and custom ERP work are scoped separately from the managed deployment.
Continue the evaluation
Continue your deployment evaluation.
No-pressure scope check
Check whether the inquiry & booking agent has a credible operating case.
The audit estimates suitability, capacity returned, setup-cost recovery, and answer confidence. A weak case should lead to a narrower scope or no deployment—not a forced sales call.
Start the Agent ROI Audit