A third of Philippine businesses now use AI, but few have a plan. Start with one workflow

AI has moved from conference talk to everyday office tools surprisingly fast. A study released in October 2026, Unlocking the Philippines' AI Potential 2026, surveyed 1,000 Philippine business leaders. Strand Partners conducted it for Amazon Web Services, a company that sells AI services, so keep that in mind. Its headline numbers are still useful.

What the study found

• 32% of Philippine businesses now use AI, up from 21% a year earlier. That's more than 396,000 businesses.
• Most use is basic. About 70% of adopters mainly use ready-made tools such as public chatbots and productivity apps. Only 12% have built custom or more advanced systems.
• Few have a plan. Only 17% have a formal roadmap for expanding AI, while 57% are still exploring or experimenting. Just 12% have a formal responsible-AI framework.
• Those using it report gains. 74% of adopters said AI improved productivity, and 68% said it increased revenue.
• Skills are the main barrier. 47% cited a shortage of AI and digital skills, followed by technical or data issues (39%) and limited budgets (36%).

In short: lots of people are trying AI, most aren't sure where it fits, and very few are managing the risks.

Why "start with one workflow" works

The businesses that get real value from AI usually don't start with a big platform. They pick one repetitive task that eats up hours, fix it, measure it, then move to the next. It's cheaper, it shows results quickly, and it teaches your team what AI is actually good at.

Good first candidates

• Drafting routine documents such as quotations, letters, job posts and meeting summaries, which a person then checks.
• Sorting and answering common inquiries, like "what are your hours" or "how do I apply", with a human taking over anything unusual.
• Turning scattered data into a daily summary, such as sales, stock levels or attendance, sent to managers automatically.
• Reading forms and receipts into a spreadsheet instead of retyping them.

Often the biggest win isn't the AI itself. It's tidying up the process around it: one form instead of three, one place for the data, and clear rules for who approves what.

A simple five-step approach

1. Pick the task. Choose one task that's repetitive, frequent and easy to measure.
2. Measure today. How long does it take per week? How many errors?
3. Try it small. Test with one person or one branch for two to four weeks.
4. Keep a human in the loop. AI drafts, people decide, especially for anything sent to clients or the public.
5. Measure again and decide. Keep it, adjust it, or drop it.

Watch out for personal data

This is where many offices slip. Pasting client names, ID numbers, health information or payroll details into a public chatbot can put you on the wrong side of the Data Privacy Act. Before anyone uses AI with real data, agree on simple rules:

• which tools are approved
• what information must never be pasted into them
• who reviews AI output before it goes out

Writing those rules down puts you ahead of the 88% of businesses that, according to the study, don't yet have a formal responsible-AI framework.

The bottom line

You don't need a big AI strategy to start. You need one well-chosen task, clear rules about data, and a way to measure the result. Do that a few times and you'll have the beginnings of a roadmap.

If you'd like help choosing that first workflow, or automating it properly, that's exactly the kind of work we do.