Yesterday I found myself searching for Media Communication programs for undergrads. After an hour of ‘research’, I had three different Top 10 lists. So I sent it off to the friend who had asked for the options. Within half an hour, I got a message: Options in Europe only.
So I spent some more time and found 10 more options and sent it off. The response was: “Europe. Not UK.”
Repeat process and I had a list of 10 again. Response: “Programs in English please.”

I was about to begin the next search but then decided to speak to some people I knew in global media. After several calls, I had seven recommendations and guess what? None of the previous 50 were among the recommendations from the media people. I looked at the new list carefully. Would this meet the requirement? Wait, what was the requirement?

So I sat back and started thinking. The past 5 hours I had been searching on Google, and then on ChatGPT, Claude and Grok. Why did I move from Google search to ChatGPT? And why did I start searching on Claude and Grok even though I was already on ChatGPT? Finally, I spoke to people who were actually in the media industry. I still wasn’t sure if this was a good list. So I got onto a call to speak to the student who required the list of college programs. 

Turns out he had already done more or less what I had done and had compiled many lists, had a lot of information and everything looked good and everything looked challenging and was therefore having difficulty in making a shortlist.
So we started over with questions… and got stuck at the very first one: What do you want to study?

“I am not sure.”
“Why did you pick Communication Media?”
“I am interested in journalism. But I don’t see what the prospects will be in the future.”
“What do the universities say?”
“They have the same kind of information. Digital media, technology, digital marketing, business intelligence, online journalism… none of that is interesting. If everything is digital, then what exactly will I be learning over three years?”
“What are you really interested in studying?”
“Many things but I don’t know if I want to make a career in those.”
“But you don’t have to make a career in what you study.”
“Isn’t that the point of going to university? To study something so that I can get a job in that field?”
“Is that why you want to go to university?”
“Isn’t that why you went to university?”
That was a very good question. Why do we go to university? How do we decide what to study? Is everybody clear about their decisions and reasons? 

When I was applying, I remember talking to everyone – friends, family, family friends, families of friends, teachers, whoever was willing to engage. Coming to think of it, I spoke to so many people because I did not find the answers I was seeking. But the boy next door had no such trouble, he was clear he was going to do engineering. The girl who lived a floor below had a great portfolio for her architecture applications. Her cousin sister who used to visit summers had taken a year off and then secured entry to a medical degree, she wanted to become a surgeon. Most of my friends had no doubts either. They all seemed to know what they wanted to do. One wanted to become an actuary and join the insurance industry. The others wanted to do a Bachelor’s in Science and then do a Masters in Technology. 

I had no idea what I wanted to do. Or rather I had too many things I wanted to do. As a result, when I thought of doing something, I used to go about finding the requirements for admission and then I would find criteria which I should have thought of a year ago, so I would shift attention to something else. Architecture, chemical engineering, statistics, journalism, design, history, physics, political science, marine biology … everything seemed so interesting and exciting. I thought of and applied for 12 different programs in 12 different universities in five different cities. I took professional aptitude tests and counselling only to find out that I could become anything I wanted to. And that didn’t help at all.

Today, I can do that research in probably a day. But I realised that such ‘research’ is not really helpful for people who are uncertain of where the future lies. A more systematic approach is required to present information to help make a decision. This is where AI platforms can be helpful. One of the advantages of Open AI’s ChatGPT is that you can configure your own custom GPT (feature available only for paid subscribers at the moment) by defining certain boundaries, a broad subject, specifying types of sources of information, conversation style, limit of responses, etc. For instance, I have just configured (no coding required, just words) a GPT to help me find information about university programs. The main purpose of the GPT is to quickly gather basic requirements (desired location, subject, interest areas within subject, scholarships, fees, types of university (new age, traditional, large or small intake), admission requirements, etc. 

(Btw, it took less than an hour to design this in collaboration with ChatGPT itself):

 Step-by-Step: Build a Custom GPT for Undecided Students

Step 1: Access the GPT Builder

1. Go to chat.openai.com
2. Click your name/profile at the bottom left.
3. Choose “Explore GPTs”
4. Click “Create a GPT” to open the GPT Builder.

Step 2: Define the Custom GPT’s Role
Prompt: “What should your Custom GPT do?”

Paste this: This GPT helps students who are undecided about what to study at university. It guides them through four broad academic categories—Interdisciplinary, Sciences, Humanities/Social, and Arts—helping them understand what each means, what subjects they include, and how to find real-time admission info based on their preferences.

Step 3: Choose Custom GPT Name & Identity
GPT Name: UniGuide GPT
Tagline: “Helping students make confident academic choices.”
Personality & Behavior:

• Friendly, warm, non-judgmental
• Thinks like a curious mentor
• Prioritizes clarity and exploration
• Uses up-to-date, reliable data via Web Browsing

Step 4: Add Custom Instructions
Prompt: “What should the GPT know about how it should behave?”
Paste this: Guide students gently through their decision-making process. Start by helping them choose from four categories: Interdisciplinary, Sciences, Humanities/Social, Arts. For each, show:

1. A short Description of what the area means
2. A list of traditional and emerging Subjects
3. A personalized Explore phase, asking them about subject interest, location, university type, and budget

Then use web browsing to return a list of up to 10 real programs with admission requirements, deadlines, fees, and links. Be empathetic and ask questions if the student is vague. Avoid assumptions and prioritize usefulness over fluff.

Step 5: Enable Tools

 Enable Web Browsing
 Code Interpreter – Optional
 File Upload – Not needed

Step 6: Test with Sample Prompts

Prompt: “I don’t know what to study, but I’m curious about combining philosophy and tech.”

Expected GPT Behavior:

• Recognize the interest fits Interdisciplinary
• Explain interdisciplinary study
• Suggest subjects like Cognitive Science, STS, or Digital Humanities
• Ask region/university type
• Fetch 10 real programs based on preferences

Why This Custom GPT Beats Search or Raw AI Chat

Search Engine	AI Search	Custom GPT
Gives pages	Gives blurbs	Guides decision
No personalization	Some summarizing	Full conversation
Overwhelming	One-shot answers	Layered, structured flow

This kind of structured Custom GPT acts like an informed buddy or coach, not just a finder. It’s particularly powerful when users are:

• Uncertain
• Not sure what the right questions are yet
• Overwhelmed by search results

Custom GPT can be an awesome way of configuring AI to be helpful for decision making in every day life. I have configured one to help students study history, one to do focused research on training for aircraft pilots, and yet another to play a never-ending quiz. But you have to be careful since AI platforms are prone to errors.

Quick Checklist When Working with Custom GPT

• AI platforms tend to provide dated (old) information at times. Double check deadlines. What I do typically is to use GPT/AI to figure out general information, make a shortlist of universities and then visit each shortlisted university’s website and get the latest info for the current academic year.
• AI sometimes overlooks key criteria. So double check outputs/lists, don’t assume it is what you needed. For instance, when I ran the exercise to shortlist Law programs in Europe for a student, I got a great list but more than half the programs were in European languages while the student was looking for English programs. I had specified this in the requirements, but the AI still included non-English programs. This happens because it is retrieving information from the internet and not necessarily fact-checking every criteria. 
• At this moment, AIs tend to provide information for queries, they don’t generally ask questions to clarify requirements. So be sure to engage and provide as much detail. If your criteria changes runtime as you search, inform the AI whether it is an overall change or only for a particular program/university. 

Study AI with Goseeko: https://www.goseeko.com/landing/international-students/
Students from UAE can enroll here https://www.goseeko.com/landing/middle-east/