Lead Capture Agent
A visitor makes an after-hours request. An approved acknowledgement gathers the details and routes them to a named person. Response windows and escalation rules are agreed with the business.
$3,000–4,500 + $300–450/mo
AI assistance for the work that piles up.
We build AI tools to organize inquiries, answer approved questions, and prepare response drafts. Your team provides the reference information and decides what needs human review. Monthly care covers errors, updates, and performance.
Setup from $2,000 · care from $200/mo · included in Growth SystemsInstrument 04/ Intelligence in motion
Put intelligence
in motion.
Try a sample inquiry and follow it through to a response draft. You’ll review the proposed handoff before the demonstration finishes.
The sample inquiry
Website inquiry · 9:40 pm
The next step
Review a sample inquiry.
The demo extracts request details, checks sample reference material, and prepares a response for review.
- Input
- A service inquiry
- AI role
- Organize and draft
- Human
- Confirm the next step
Human review before action
Connected
ScatterConnect
Interactive demonstration. Sample messages and a simulated workflow. Nothing is sent, booked, or connected to a live AI service.
Bennu Lead Agent
9:40 pm · after hours
Scripted demonstration of a proposed workflow. No message, CRM update, or appointment is created by this example.
what this actually looks like
An after-hours inquiry example.
This scripted example shows an after-hours inquiry being collected and routed to a team member, who checks availability and confirms an appointment.
- Pricing is handled by your team.
- The assistant identifies itself as automated.
- Service area and hours come from approved information.
- A team member confirms the appointment.
The transcript is a demonstration, not a record of a customer inquiry or a measured result.
six named use cases
Six uses we can assess.
Each use case has a one-time setup fee and required monthly care. We confirm the scope, reference material, and review process before building.
FAQ / Knowledge Agent
Answers routine questions from an approved FAQ and refers questions outside it to your team. Unanswered questions can help identify gaps in the reference material.
$2,000–3,500 + $200–350/mo
Follow-Up Assistant
Prepares follow-up drafts for unanswered quotes using the account record and approved language. A sales representative reviews the message and decides when to send it.
$3,000–5,000 + $300–450/mo
Review Response Drafts
Prepares a response draft to a new Google review within a day. You review and approve it before posting.
$2,000–3,000 + $200–300/mo
Internal SOP Agent
Helps staff find information in approved procedures, with source references they can check. Technical and safety decisions remain with qualified people.
$2,500–5,000 + $250–400/mo
Quote Intake / Triage Agent
Extracts product, quantity, location, and timing details from quote requests, then routes a summary to sales. Acknowledgement rules are agreed in advance; people prepare the prices.
$3,500–5,500 + $350–500/mo
the boundaries
What AI should not do.
These limits are part of how we scope and review an AI project.
- Give pricing, legal, medical, or safety answers it improvised.
- Present itself as a person. The system identifies itself as automated.
- Send anything involving money, commitments, or complaints without human review.
- Take responsibility for decisions that belong to your staff.
- Operate without monitoring and review.
We’ll identify any process or information gaps that need attention before a build.
- Business-tier platforms only. Your customer data never goes into consumer tools.
- Agents answer from your approved knowledge — and decline outside it.
- Human approval remains required for pricing, commitments, and other high-stakes decisions. Customer-facing drafts start with human review; any automatic acknowledgements use an explicitly approved scope.
- Defined escalation rules route requests to a designated person.
- You own your data and transcripts. Every build ships with a one-page data-flow map.
- Monthly care includes error review, reference updates, and a performance report.
- Your prices, hours, and services change. The agent’s knowledge has to change with them.
- Regular review helps identify errors and changes needed in the reference material or instructions.
- Monthly care costs $200–500, depending on the system, and is required with every build.
straight answers
Questions clients ask
Where does our data go?
We use business-tier platforms with appropriate data agreements. The project includes a one-page diagram showing which services receive the data. You own your data and transcripts; customer information is not entered into consumer chatbot accounts.
How do we know it won’t say something wrong to a customer?
We constrain the system to approved information, test it against representative questions, and define when it must hand off to a person. Customer-facing drafts begin with human review during a 60-day supervised period. Only explicitly approved acknowledgements may send automatically; pricing and commitments continue to require human approval. Ongoing care includes review of errors and needed updates.
Will this replace anyone on my staff?
The scope is to assist your staff with repetitive work. People remain responsible for pricing, commitments, complaints, and the review of AI output.
What if AI isn’t right for our business yet?
We may recommend improving a process or its reference information first. The $1,950 audit assesses 3–5 possible uses, including estimated cost, potential benefit, and any reasons to wait.
How long does a build take?
A focused system is typically scoped for two to four weeks to an initial supervised launch, followed by a 60-day supervised period. Timing depends on the approved use case, access, reference material, and integrations; we confirm it in the proposal.
the right question
Have a repetitive task in mind?
Tell us what takes up your team’s time. We can discuss the task or assess it as part of the $1,950 audit, which includes 3–5 potential AI uses and their costs, assumptions, and limitations.