Using an AI agent as my personal assistant

I HAVE always wanted a personal assistant but it would be a luxury. Thanks to artificial intelligence, it is now possible to have more than one personal assistant working for you.

I was researching Iceland for a self-drive family vacation next year. The exercise was daunting because there are so many permutations in terms of driving routes, places to stay and visit. After seven versions of the itinerary, I caved in and subscribed to ChatGPT Plus for assistance.

To justify paying S$28.98 monthly, I decided to leverage the functions and make it work harder for me.

Life became a lot easier

An AI agent makes parts of that personal assistant support accessible. This includes researching holidays, preparing financial comparisons and organising health records. For me, its appeal lies in reducing the work between having a goal and making informed decisions.

Instead of repeatedly searching, copying information and updating spreadsheets, I can give my AI agent a clear brief and review the result.

An AI agent combines conversation with tools that can carry out tasks. Depending on the service, enabled connections and permissions, it can browse websites, examine documents, develop presentation slides and edit files very quickly. This makes it useful for work involving several steps.

My role is to set priorities, provide relevant information and decide which actions to take. The AI agent’s role is to gather evidence, organise options, highlight uncertainties and monitor progress.

Planning for travel

Travel was my starting point. A holiday involves many connected decisions: flights, accommodation, routes, attractions, meals and transport. Changing one overnight stop can affect several days. An AI assistant can maintain a consolidated itinerary and revise the surrounding arrangements whenever I change my plans.

For my Iceland road trip, my priority is to keep daily driving manageable. It could compare alternative overnight stops, estimate travel times, and allow space for sightseeing, meals and weather disruptions. The itinerary distinguished confirmed bookings from suggestions, and identified information that still needed to be checked.

For another itinerary – this time a trip with extended family to Hainan – the AI agent compared taxis with a hired vehicle. It considered passenger numbers, luggage, convenience and the cost per person. It also grouped nearby attractions and restaurants to reduce unnecessary journeys.

I learnt to provide detailed instructions for better results. Before departure, I could request a fresh check of opening hours, transport schedules and official travel notices. Of course, I would approve bookings and payments myself.

Regular reports on passive income

Financial research is another area where an AI assistant could save me time. Rather than collecting promotional material, I requested a weekly report on passive income options and comparisons relevant to a Singapore resident.

The brief covered Singapore Savings Bonds, Treasury bills, fixed deposits, income funds, dividend shares and annuities, where relevant to my circumstances. Each report should use current official information and show its research date. It should explain minimum commitments, fees, access to capital, payout frequency and the risks behind the advertised return.

Most importantly, the assistant should distinguish interest, investment distributions and withdrawals of my own capital. A high payout is not enough to establish that an investment meets my needs.

I could ask: “Compare these options against my desired monthly income, emergency cash needs and tolerance for losses. Explain what could go wrong.” Its report would help me prepare questions for a qualified adviser while investment decisions would remain mine.

Simplifying retirement insurance

Retirement insurance brochures can be difficult to compare because products use different premium periods, payout structures and benefit illustrations. I would upload brochures and quotations, then ask the assistant to arrange their features in a consistent table.

The comparison should separate guaranteed benefits from projected bonuses, show when premiums and payouts occur, and explain surrender terms.

I would also ask the assistant to identify gaps: Does income last for life or a fixed period? How much flexibility remains if my circumstances change? What happens if I stop paying premiums? Any calculations should state their assumptions and use the actual cash flow dates.

Daily fitness and nutrition companion

Health tracking is likely to be used more frequently. I send daily entries for weight, body fat, muscle mass, biological age and steps. AI estimates the calories for breakfast, lunch, dinner and snacks. With access to my tracker, the assistant can update the spreadsheet, calculate totals and identify missing entries.

Food photographs make logging easier, especially when eating out. However, calorie estimates should show a range and assumptions about portions, oil, sauces and drinks. If I ate only a little rice or shared several dishes, that information should change the estimate.

Weekly reviews could turn the log into useful feedback. I would ask for average steps, weight trends and recurring eating patterns, alongside manageable suggestions.

The assistant should treat device readings cautiously and avoid diagnosing conditions or setting restrictive targets without appropriate context. Persistent symptoms or concerns should be channelled to a healthcare professional.

Building a routine

An AI agent’s greatest value would be helping me follow through consistently and efficiently. After all, delegating with clear instructions is not new to me. I would still choose where to travel, how to invest and what health changes to make, with more organised information supporting those choices.

Having an AI agent do the heavy lifting helps to enhance my financial and physical wellness, leaving me more time to travel and enjoy life.

Note: As a safeguard, never input real personally identifiable information or sensitive data into an AI tool.

The writer, an IBF Fellow for financial planning, is author and head of financial planning literacy at DBS