Meet Salesy, Jobby, and Penny: Build Your AI Back Office
Give AI helpers clear roles in sales, dispatch, and billing, with example prompts and reports built around your home-service business.
· Illustrative example
Not hands-on tested. These are proposed workflows, not verified native Housecall Pro integrations. Confirm current product access, account requirements, and connection behavior before use.
Your business already has departments, even when the same person handles three of them.
Someone checks estimates. Someone organizes tomorrow's jobs. Someone makes sure invoices get reviewed. AI assistants can help prepare the information for those jobs when you give them a clear role and a reliable source.
Here is a fictional example for a home-service company.
An illustrative AI back office: Salesy reviews estimates, Jobby briefs the team on jobs, and Penny flags invoice questions.
Original Data Flow Pros graphic; fictional example.
Fictional figures: Salesy flags three estimates totaling $8,450 in quoted value; Jobby lists 12 appointments, two needing a technician; Penny flags four invoices totaling $3,180 past due. Quoted value is not booked revenue. Agent names describe roles, not separate accounts or guaranteed permission boundaries.
| Helper | Job | First useful output | Human decision |
|---|---|---|---|
| Salesy | Review estimate opportunities | A list of sent estimates worth checking | Whether to contact the customer and what to offer |
| Jobby | Prepare the daily job brief | Appointment windows and missing assignments | Changes to dispatch and technician assignments |
| Penny | Flag billing items for review | Invoices matching an agreed due-date rule | Collection messages, corrections, and payment actions |
You can begin with one assistant handling three named responsibilities. Create separate Bots or agents only where your platform supports them and there is a useful reason to do so. Names help organize the work; permissions and data access need their own setup.
Give every helper a short job description
Write down five things: what it reads, what it produces, when it works, who checks it, and what it is allowed to change.
Salesy's first job might be to prepare estimate candidates each Monday. Jobby's might be to summarize tomorrow's appointments. Penny's might be to review invoices every Friday. These are different responsibilities even if one assistant performs all three.
Fictional Salesy report: three estimates worth $8,450 are flagged for review, with missing contact history explicitly identified.
Original Data Flow Pros example with invented records, not customer data or a product screenshot.
Quote value is not booked revenue. These are candidates for human review, not proof that no follow-up happened.
Teach the business rules
Explain your service area, working hours, approved wording, status definitions, and escalation rules. Business instructions explain how to interpret the work. A live connection supplies the records. Uploading a company handbook does not create live access to Housecall Pro.
Give the work a home
Start with private reports. When ready, choose a destination your team uses, such as a supported Slack or Teams connection or an automation that delivers to an approved channel. Verify both the destination and permissions before posting. A tool's ability to read Housecall Pro and its ability to send a message are separate capabilities.
Measure something small
For the first month, track how many flags were accurate and whether the office used the report. Fix unclear definitions before expanding the workflow.
Start with one helper and one job your office repeats every week. Data Flow Pros can help connect the information and build from there.
Source context: Sintra Brain AI illustrates shared business knowledge; Grok Bot documents named Bots. The example roles and operating model above are Data Flow Pros editorial proposals.
