Admin & Operations
The Business Owner's Guide to Saving 5 Hours a Week With AI (A Real Example)
Australian small business research found that admin costs owners the equivalent of 13 months of work — inside a 12-month year. One full month, lost to emails, invoices, scheduling, and paperwork. Sarah (name changed, story composite) is a sole-trader bookkeeper in Melbourne who started using ChatGPT three months ago. She doesn’t describe herself as tech-savvy. She’s not automating anything. She’s just using AI the way she used to use a notepad — to think out loud and get things done faster.
Who Sarah Is
Sarah has eight to ten active clients at any given time — mostly small businesses in Melbourne’s inner north. She’s been in practice for six years, works four-and-a-half days per week, and does everything herself: client work, admin, marketing, scheduling, and invoicing.
She tried AI twice before. “Both times it felt like talking to a robot,” she says. “I asked it something specific and got back something generic.” The third time, a bookkeeper friend showed her the 4-part prompt formula — give it your role, the context, the task, and the format you want. “That changed everything. It stopped feeling like a guessing game.”
This is her week.
Monday: Email
Sarah starts Monday with 11 unread emails needing a reply. A mix of client questions, scheduling requests, and one potentially awkward message from a client unhappy about something small.
Before AI: 45 to 60 minutes of staring, drafting, editing, re-reading, second-guessing.
With AI: a 3-line prompt per email for the straightforward ones. For a batch of three client enquiries asking similar questions, she pastes all three into one session: “Here are three client enquiries. Draft professional responses for each.” Done in 8 minutes.
One email gets individual attention: a client asking to reschedule their quarterly review — the third time in two months. Sarah wants to be warm but honest about the impact. She types:
Write an email to a bookkeeping client who is asking to reschedule
their quarterly review for the third time. Tone: warm but gently
honest — let them know it's fine, but that regular reviews help
us catch issues early. Keep it friendly, not passive-aggressive.
Under 100 words.The draft comes back measured and kind. She reads it, adds one line specific to their situation, and sends. Four minutes total.
Time saved Monday: ~40 minutes.
Tuesday: Invoicing
End of month falls on Tuesday. Sarah has eight invoices to prepare.
Before AI: line item descriptions — the “what did I actually do for this client in plain English” field — took her 5 to 8 minutes each. Total: up to an hour just writing descriptions.
With AI, she runs one batch prompt. She opens ChatGPT and types:
I'm a bookkeeper in Melbourne preparing end-of-month invoices.
I need professional line item descriptions for 8 client jobs.
Keep each under 40 words, specific, and professional. Here are the details:
Client 1: Monthly reconciliation and BAS preparation for a retail client.
June quarter.
Client 2: Catch-up bookkeeping for a new client — 3 months of transactions
entered and categorised.
[etc. for all 8 clients]Eight descriptions come back in one go. She reviews them, adjusts two where she wants more specific wording, and she’s done in 14 minutes.
Time saved Tuesday: ~45 minutes.
Wednesday: FAQ Responses and New Client Enquiries
Three new enquiries arrive through her website contact form. All three are variations of questions she’s answered dozens of times: “How do you charge for BAS preparation?” “Do you work with tradespeople?” “What software do you use?”
Before AI: she typed fresh answers each time. Good answers, usually — but starting from scratch every time.
With AI: she opens her FAQ response Google Doc (built over the past month using the technique from this series). She finds the nearest prompt template, pastes it in, tweaks the details for this specific enquirer’s context, and sends.
All three done in 11 minutes.
One of them converts to a discovery call the same afternoon. Sarah thinks the response helped — it was detailed, warm, and specific. “It would have been fine before. But it was better this time.”
Time saved Wednesday: ~30 minutes.
Thursday: Document Review
A new supplier contract arrives. Software subscription for her client portal — 18 pages of terms, data handling clauses, pricing conditions, and renewal terms.
Before AI: she would have read it over two sittings, taken some notes, probably missed something.
With AI: she pastes the full document into Claude — she uses Claude for longer documents because it handles more text at once. She asks:
Summarise this software subscription agreement. I'm a small business
owner in professional services. I want to know:
1. Payment terms and any price increase clauses
2. Cancellation notice requirements
3. Data ownership — who owns my client data?
4. Automatic renewal terms
5. Anything unusual I should ask about before signing
Plain English. Bullet points.She gets a 5-bullet summary in 40 seconds. Two clauses are flagged as worth checking — she reads those sections in the original, calls the supplier with one question, and is comfortable signing.
“I’m not skipping the contract,” she says. “I’m reading the right parts.”
Time saved Thursday: ~50 minutes.
Friday: The Calendly Update
Three months ago, Sarah’s Calendly booking page copy was written in a rush on a Sunday night. It’s been bothering her since — vague, slightly impersonal, not quite what she’d want a new client to read.
She’d been putting it off because it felt like a task that deserved a proper hour, and she never had a proper hour spare.
Friday afternoon, 15 minutes before she closes for the week, she runs a prompt:
Write copy for a bookkeeper's Calendly booking page for a
"15-minute discovery call." The message should:
welcome the visitor, briefly explain what happens in the call,
list what to have ready (basic info about their business,
any specific pain points they're hoping to solve),
and end warmly.
Business: sole-trader bookkeeper in Melbourne specialising
in small businesses. Tone: professional and warm, no corporate language.
Under 150 words.She reads the result, makes three edits, and publishes it to Calendly.
“That had been on my to-do list since January.”
Time saved Friday: ~25 minutes — plus three months of procrastination.
The Weekly Tally
Add it up:
- Monday (email): 40 minutes
- Tuesday (invoicing): 45 minutes
- Wednesday (FAQ responses): 30 minutes
- Thursday (document review): 50 minutes
- Friday (Calendly copy): 25 minutes
Total: just under 3.5 hours that week.
That wasn’t her best week. Some weeks it’s closer to 5 hours — when there are more invoices, harder emails, or documents to review. What she reclaimed on Thursday afternoon: two new-client discovery calls she’d been scheduling into next week. One converted.
ChatGPT costs her $30 a month. “I made that back on Monday morning,” she says.
The Real Point
Sarah isn’t automating her client relationships. Her clients still get her expertise, her judgment, her familiarity with their books. What they don’t get is the version of her that’s exhausted from admin.
She uses AI the way she used to use a template — to start faster, not to think less. The thinking is still hers. The writing is faster. The week is shorter.
“I didn’t get into bookkeeping to spend my evenings writing emails,” she says. “AI gave me the evenings back.”
If you want to know which three tasks in YOUR specific business AI can handle this week, that’s exactly what the free chat below is for.
Book Your Free AI Audit
💬 Free 30-Minute AI Time-Saving Audit
Want to know which 3 tasks in YOUR business AI can handle this week? Book a free 30-minute call — I’ll look at how you currently spend your time and show you exactly where AI fits. No pitch. Just a practical plan.
[Book your free audit call →]