You will build a one-page inventory of the AI in your working life, sorted by layer and by risk.
Most people, asked how much AI they use at work, name one chatbot and stop there. Ask the same people to walk through a normal Tuesday and the count climbs fast: the email filter, the autocomplete in Outlook, the meeting transcript, the translation button in WhatsApp, the dashboard that flags overdue invoices. This exercise turns that Tuesday into a one-page inventory, so you know where AI already touches your work and where a mistake from it would hurt.
You need about twenty minutes, the inventory table template below this lesson, and your calendar or sent folder from last week to jog your memory.
Go through a normal working week day by day. For each tool you opened, ask one question: does it make a prediction, suggest something, or produce content for me? If yes, it goes on the list.
Be generous at this stage. Include the small things: the "suggested reply" chips in your email, the spellchecker that rewrites a whole phrase, the HR system that ranks job applicants, the calendar that proposes a meeting time, the bank app that sorts your spending into categories. Include tools your company runs for you even if you never see them, such as the filter that quarantines suspicious attachments. Aim for at least eight rows. Most people find twelve or more once they look.
Next to each tool, write one of three labels, using the layers from lesson 1.1, AI is a family of techniques, not one thing.
Rule-based means a person wrote the logic in advance, like a form that rejects a claim over a fixed limit. Machine learning means the tool learned a pattern from past examples and gives a label, a score or a ranking, like a spam filter or a recommendation list. Generative means it produces new text, images or audio, like a chatbot, a meeting summariser or an image tool.
You will not be sure about some of them, and that is fine. Mark those with a question mark rather than guessing. Vendors rarely say how a feature works, and a question mark is honest. It also tells you which tools to ask your IT team about.
For every generative tool, add two short notes. What do you check before you use its output? What do you not check?
Write down what you actually do on a busy day. If you paste the meeting summary into an email to your boss without reading it against your own notes, write that down. The point is to see your real habits. A generative tool fails by producing something fluent and wrong, so the gap between what you check and what you skip is where its errors reach other people.
Read down the list and ask, for each row, what happens if this tool gets it wrong and nobody notices. A spellchecker that changes "their" to "there" costs you a little embarrassment. A summary that drops a client's deadline, or a screening tool that filters out a strong candidate, costs a lot more.
Circle the single tool whose unnoticed mistake would cost you most, in money, reputation or someone else's welfare. Modules 6 and 8 will keep returning to that tool, so choose honestly.
Here is part of the inventory that Nurul, an HR executive at a logistics company in Jurong, filled in. Treat her tools and habits as an example of the method, since nobody is suggesting you copy them.
Applicant tracking system ranks CVs for each opening: machine learning, with a question mark because the vendor only says "smart matching". She reads the top twenty and never opens the rest. Outlook suggested replies: machine learning. Low stakes. Meeting transcription in Teams: generative for the summary, with the speech-to-text part being machine learning. She checks names and dates in the summary but not the action items. Leave system that blocks requests beyond the balance: rule-based. ChatGPT for drafting job descriptions: generative. She checks the salary range and the job title, but not the list of duties, which she has noticed it pads out. Payroll software that calculates CPF contributions: rule-based. Google Translate for messages from a foreign worker's family: generative, since it writes new text in another language. She does not check it at all because she cannot read the language. LinkedIn's suggested candidates: machine learning.
Nurul circled the applicant tracking system. A ranking she cannot see inside decides which CVs a person ever reads, and she never looks below the top twenty. The translation tool came a close second, because an error there would reach a family that could not tell her it was wrong.
When your own table is filled in, read it from top to bottom once more. Most people notice two things: they use more AI than they thought, and the tool they trust most is often one they never check. Keep the finished page somewhere you can find it, because you will compare it with your answers at the end of the course. Open the template now and start with Monday.
Complete the tool inventory table with at least eight rows and write two sentences on which tool you trust most and least, and why.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).