You will be able to name the personal finance tasks where an assistant saves you real time.
Asking an AI assistant to "help me with my finances" usually gets you advice that could apply to anyone: spend less than you earn, consider index funds (funds that track a market index, such as the STI). The assistant still has its uses. A request that open gives it a job it is bad at, which is knowing your situation and telling you what to do.
Lesson 1.1, "What an assistant is actually doing when you ask about money", showed that a language model is strong with words and structure and weak with current facts and arithmetic. Four money jobs use the strong side and save real time: sorting, explaining, drafting, and planning what to check.
Darren, an invented example, is 29 and works in operations at a logistics firm in Jurong. He has meant to sort out his money for about two years. Last Sunday he opened an assistant, typed "help me with my finances", got eight hundred words of generic advice about spending less than he earns and considering index funds, and closed the tab.
Bank and card statements are often hard to read. Merchant names get cut off after twelve characters, some lines are strings of payment codes, and many shop names say nothing about what the shop sells.
PayNow lines name a person, not a shop. When someone pays you by PayNow, your statement shows their full account name, which may not be the name you know them by. Since 6 June 2026, the masked name with some letters hidden is what a payer sees for an individual payee while making a transfer, not what appears on a statement. Your bank's PayNow FAQ explains how it displays these names. Sorting is handing those lines to the assistant and asking which category each probably belongs to: food, transport, bills or the rest.
This is slow work for a person and a good fit for a language model. It has seen huge amounts of text about shops, brands and payment services, so it can usually tell that a truncated name is a supermarket, or that a ride-hailing code goes under transport. Two hundred rows take it seconds and would take you an evening.
The question you are asking is what kind of thing a messy label probably is, which is a language task. You are not asking it how much you spent. It will make mistakes, and Module 2 shows you how to catch them.
Darren's company recently changed its group insurance, and the summary uses three terms. The deductible is the part of a bill you pay before the insurer pays anything. Co-insurance is the share of the rest that you still pay. The pro-ration factor is a cut to your claim if you stay in a costlier ward than your plan covers. His investment app shows an expense ratio on every fund, which is the yearly fee the fund takes as a percentage of your money. He has a rough idea what each term means, which is the dangerous level of understanding.
Explaining is asking the assistant to turn jargon into plain words. It does this well, and better still when it gives a worked example with made-up numbers. Ask it to explain a term "as if to a friend" with a concrete example, and the answer is usually clearer than the policy wording. For a deductible, it might walk through a S$5,000 claim (an example figure) and show what you pay and what the insurer pays.
The assistant gives you the common meaning of a term, and your own policy may define the word differently. Module 3 deals with that. For learning what a term generally means, an assistant is fast, patient and does not mind being asked the same thing several ways.
Many money tasks are really writing tasks: a polite but firm email to your bank about an unexpected fee, a list of questions for an insurance agent, a message to your landlord about your deposit, a checklist before you sign up for a gym package or a renovation contract. Most people already use assistants for writing at work, and the skill carries over.
Drafting divides the work. You supply the facts, the assistant supplies structure and tone, and you edit the result. Every fact in the draft has to come from you. If the assistant adds a figure, a policy term or a deadline you did not give it, delete it or check it before you send anything.
Planning what to check is the job people miss. Before a decision you need a list of the facts that matter, such as what the product costs each year, how you get your money out, what happens if you stop paying, and who regulates the provider. Lists of considerations are a language pattern an assistant has seen many times, so it produces them well.
Ask it for the list and find the answers yourself. Asking what to find out before choosing a savings plan gets you a useful checklist. Asking which savings plan to choose gets you a confident opinion built on facts that may be old or invented. Module 4 builds this job into a full method for researching an investment.
Darren could have used it on Sunday. Asking for six questions to put to his HR team about the new group plan would have given him a sharper conversation with them.
Each job works on language you already have or can supply:
Sorting reads labels you already have and suggests what they are. Explaining rewords ideas already written down somewhere. Drafting arranges facts you supply into clear writing. Planning lists questions without answering them.
None of them makes the assistant the source of a number you will act on, asks it for a current rate, fee, cap or total, or asks it to decide for you. Those missing jobs are the ones Darren gave it, and they are the subject of Lesson 1.3, "Where it fails: invented numbers, old facts, wrong country".
Before you get there, look back over your own last month. Most people find that several of the money tasks they did, or put off, fit one of these four jobs, and a few fit none of them.
Write down five money tasks you did in the last month. Next to each, mark which of the four jobs, if any, an assistant could have helped with.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).