Office automations with a paper trail.
I set up four specific automations for small offices of 1-10: request intake and follow-up, shared-inbox triage, new-client onboarding, and document-to-spreadsheet capture. Each is built in your own accounts and uses AI only where it helps: classifying, extracting, drafting. A person stays in charge of anything your customers see. Price is fixed and agreed before work starts. You get a written handover: process map, data-flow sheet, review rules and an off switch. We measure time or response speed before and after.
Request help and get your price
The four automations
1. Request intake and follow-up
A form or email becomes a record in a list or CRM, an immediate acknowledgement, a task and a reminder if nobody replies. AI classifies the request and drafts the acknowledgement; a person sends anything beyond the standard acknowledgement.
2. Shared-inbox triage
Mail to a shared inbox gets labelled and prioritized, with a draft reply drawn from your approved answers. AI classifies and drafts; a person reviews every draft and nothing sends itself.
3. New-client onboarding
An accepted quote or contract produces the folders, intake form, welcome email, checklist and scheduling invite. An optional AI summary reads the intake answers for you.
4. Document-to-spreadsheet capture
Invoices or forms in PDF or email are turned into fields in a sheet or accounting import. Doubtful cases go to a review queue instead of a wrong cell.
How an automation is delivered
- Free quote with five questions: weekly volume, current tools, who is involved, which personal data it touches, how long it takes today.
- Baseline measured and price approved in writing.
- Built in your own accounts; you pay tool subscriptions directly to the vendor.
- Tested on real or anonymized samples.
- Handover in 30 to 45 minutes.
- 30-day adjustment window.
- Before/after review, optional monthly monitoring.
What you get in writing
- One-page process map.
- Data-flow sheet: which personal information touches which service, where it is stored, AI provider settings, who has access.
- Human-review rules.
- Runbook: how to pause or switch off, what to check, who to call.
- Access register: accounts, connectors, owner, where credentials are stored (never the passwords themselves).
- Change log and before/after sheet.
Limits
- No autonomous replies to your customers.
- No health or regulated data at the start.
- No decisions about people (hiring, credit).
- No training of custom models.
- Ongoing monitoring is a separate agreement.
- Changes after the 30-day window are billed at the menu rate.
Price
Fixed and agreed upfront, like every project. No prices are listed on this page because no two offices automate the same volume; the free quote is where the number comes from.
Proof, honestly
No case studies yet. What I can show you today is the handover package above, the written-approval rule and a measured before/after on your own workflow. Real cases will appear here only with the client's written permission and a before/after number.
Questions
Do I keep control?
Yes. AI only classifies, extracts or drafts. A person reviews anything beyond the standard acknowledgement, and nothing sends itself to your customers.
Where does my data go?
Only through the services the automation is built on, in your own accounts. You get a data-flow sheet naming each service, where it stores data and who has access.
What if it breaks?
You get a runbook: how to pause or switch the automation off, what to check and who to call. Changes after the 30-day adjustment window are billed at the menu rate.
Can you switch it off?
Yes. Every automation has an off switch, documented in the runbook, and the underlying steps keep working manually while it is off.
Request help and get your price
Last updated: November 21, 2026