Ai for business
5 repeat tasks every Davao business can automate
This article lists five repeat tasks that most small businesses in Davao have, and shows how each one can be handed to AI for a first draft. You will also get a simple test for picking which task to start with, and what to check before you hand over any data.
Claude and Claude Code are Anthropic's products. I am an independent consultant. I set Claude Code up for businesses, and I am not affiliated with Anthropic.
What "automate" means here
I do not mean a robot that runs your business. I mean this: the AI does the first pass, and a person checks it.
That rule is the same for all five tasks. The AI drafts, and a person approves. Anything that sends, pays or deletes waits for a person. If you want to know what Claude Code is first, read What Claude Code can do for a small business in the Philippines.
Task 1: Weekly and monthly reports
What it looks like now. Someone opens the sales file, the inventory file and the bank export. They copy numbers into one sheet, fix the layout, and send it to the owner. It takes part of a day, and it comes around again next week.
What changes. Claude Code reads the same sources each time and builds the report on a schedule. The owner or the staff member opens a finished draft and checks the numbers that matter.
A Davao picture. Imagine a small trading business that sells to sari-sari stores around the city. Every Monday the office builds a list of what sold, what is running low, and who has not paid. That list follows the same layout every week. It is a good fit.
Task 2: Invoices, forms and contracts
What it looks like now. Supplier invoices arrive as PDFs, phone photos or email attachments. A staff member types the date, the supplier, the amount and the reference into a sheet or an accounting system.
What changes. Claude Code reads the document and places each detail where it belongs. It can also flag lines that look wrong, such as a missing date or a name that does not match your records. Your staff check the flagged lines instead of typing every line.
A Davao picture. Imagine a small accounting office that handles invoices for many clients, and each client sends them in a different way. Reading and sorting them is the repeat work. Checking them is still the accountant's job.
Task 3: Customer messages
What it looks like now. Customers ask the same things on Facebook, email and your website: opening hours, services, how to book, where the branch is. Your staff know the answers by heart, but replies are slow when the team is busy.
What changes. Claude Code drafts the reply using your own answers and tone. Your staff read it, change it if needed, and press send. The AI does not send on its own.
A Davao picture. Imagine a clinic or a real estate agent that gets the same inquiry many times a week. The draft is ready when the staff member opens the inbox.
Task 4: Follow-ups and reminders
What it looks like now. Someone chases quotes, unpaid invoices, appointments or deliveries by hand. The list of who to chase lives in a sheet or a notebook. When the day gets busy, people are missed.
What changes. Claude Code reads the list, finds who is due for a follow-up, and drafts a message for each one. A person reads the drafts and approves them. Nothing goes out until they do.
A Davao picture. Imagine a small logistics office with parcels waiting on a customer reply, and customers asking "where is my parcel?" The drafts are prepared from the delivery sheet. Staff send the ones that look right.
Task 5: Tracking and clean-up
What it looks like now. The business runs on one shared spreadsheet that is slow, messy and fragile. Staff enter the same information in more than one place. Someone spends part of every week merging and fixing records.
What changes. Claude Code can help build a small tracker, a simple dashboard or a form that fits how your team really works. It can also find duplicates and flag records that do not match, for a person to confirm before anything is changed or deleted.
A Davao picture. Imagine a shop with two branches that keeps stock counts in two separate files. One tracker, with a drafted weekly check for mismatches, is easier on everyone.
A simple test: which task first?
Pick the task that has all three of these:
- It happens every week or every day.
- It follows rules a new staff member could learn in an afternoon.
- A person can catch a mistake in a minute or two, before it matters.
Reports, invoice entry and common customer questions usually pass. Work that needs judgment, trust or a personal relationship usually does not. The AI can still help with a first draft there, but a person leads.
I made a free checklist for this. It has tick boxes for each of the five tasks above and a score at the end. You can get it with the form at the end of this article.
Before you hand over any data
Some of these tasks involve personal information: customer names, patient records, staff files. In the Philippines, the Data Privacy Act of 2012 (RA 10173) applies. I am not a lawyer, so ask a lawyer about your own case.
In my work, the rules are the same each time:
- An NDA is signed before I see your data.
- The AI gets the least access it needs, starting with read-only.
- A person approves anything that sends, pays or deletes.
I explain each of these in Is my data safe with AI?.
What I do not promise
I do not promise how many hours any of these tasks will save before I have seen your work. Some tasks are not worth automating. That is why a setup starts with a paid discovery week. I map the work, measure how long it takes, and give you a written plan with the hours it could save. If the numbers do not work, you stop there.
What to do next
- Write down the three tasks your team repeats most often. Note how long each takes in a normal week.
- Run each one through the three-part test above.
- If one or two pass, see what a setup covers in What a Claude Code setup includes, then apply for a discovery week.
I read every application, and the next step is a short call. You do not need to decide what to automate before you apply.