Pakistan doesn't have career counsellors. So she built an AI one.

A platform built because nothing else existed

The episode opens with Muzamil acknowledging something he has been saying on the show for a while: career counselling is one of the gaping holes in Pakistan’s education system. Viewers keep writing in, asking him directly which degree to take, which university to pick, whether data science is better than electrical engineering. He answers what he can, but he is one person with one set of experiences. “I can only guide you as good as my own experience,” he tells Dr. Zunaira Saqib early on. The episode is the result of an email he received about a platform trying to fix exactly that gap.

His guest is Dr. Zunaira Saqib — faculty at NUST Business School in the human resource department, PhD from the University of Leicester, double masters including an MSc in HR from Lancaster University, and the founder of Merafuture.pk, which she describes as Pakistan’s first AI-based career counselling company. Muzamil’s framing is plain: he has nieces and nephews in the United States who routinely take aptitude tests to figure out where their interests sit. In Pakistan, that scaffolding does not exist.

A confused kid who memorised the maths book

Zunaira’s own path is the first thing Muzamil asks about, and her answer is the kind of unguarded story the show is built for. She grew up in Lahore, studied at her mother’s school — “it wasn’t a very good experience but anyway there wasn’t a lot of choice available” — and was told to take sciences because of scope, despite the fact that maths and the sciences were not where her mind worked. She is direct about how she got through: “I will tell you this honestly and this is not a lie — the entire maths book, I memorised it. The whole book.” Maths was taught in Urdu at the time, which she notes made it worse. She ended up with eighty-nine marks out of a hundred and still cannot, to this day, solve the equations.

She names this as a flaw in the system rather than a flaw in herself, and Muzamil agrees. From there she lost both her parents within a six-month span — what little guidance she had disappeared — and her career became, by her own description, a sequence of experiments. Fine arts. English literature. Even a Punjabi subject at one point. Computer sciences for an undergrad. Then the pivotal moment: an MBA she actually loved, which is where marketing and HR found her. Then an MSc in HR at Lancaster. Then teaching at NUST. Then a PhD, finished in 2018. After that she decided she had given academia what she had to give and wanted to contribute somewhere else. Career counselling was that somewhere else.

Why human counsellors can’t solve this

Muzamil makes a structural point about halfway in, and it lands hard. Sindh province alone has roughly forty to forty-five thousand schools. Even if he decided today to train forty thousand career counsellors, he doesn’t know what degree they would even hold — career counselling is not a credentialled field in Pakistan. Punjab is larger again. Zunaira’s response is the entire wedge for her company: “It’s very unpractical. Governments can work on it. But we need a more disruptive solution.”

Her solution is to invert the usual aptitude-test logic. Most testing in Pakistan, she argues, is subjective. A child has a cousin who became a doctor, or a teacher who suggested engineering, or a parent who watched a neighbour’s son do well in finance. Even when a formal test exists, there is rarely an objective model behind it — it produces a result without a basis. Merafuture spent its first year doing aptitude tests on people already considered successful in different careers. Computer sciences. Business administration. Natural sciences. Roughly eight to ten broad career fields available in Pakistan. The patterns from those successful cases became the training data for the company’s AI model.

A child who comes to the website now takes a forty- to fifty-minute test. The output is not just a personality and interest profile. It is a match against the successful-cases model — a statement that, for example, the student’s subject area and personality matches a successful business administration professional at roughly eighty percent. Zunaira is emphatic about why this matters: the recommendation is objective, no human involved, generated by matching the student against people already doing well in the field. The model is self-learning, updated every year, because — as she puts it — what counts as success today will not be what counts as success four years from now.

The pricing decision that defines the company

When Muzamil asks about the financial model, the answer doubles as a mission statement. Tests started at a thousand rupees per person. They have since moved to fifteen hundred, because, as Zunaira notes, the business has to survive. When Merafuture works with schools and colleges, the price drops. When it works with non-profits — she names Tameer-e-Millat, the Hawwa Foundation, and the Aga Khan Foundation — it operates on a no-profit basis. The platform has served roughly ten thousand students so far, slowed considerably by the pandemic closing schools. Her five-year target is sixty-five thousand students per year on the site.

The deliberate cheapness is not just pragmatic. She returns to it repeatedly: “Are we becoming the company which is serving the elites? This is something I do not want to do.” It is the reason she resists turning the mentorship layer into a premium product, because the moment industry mentors who charge five or six thousand rupees an hour are introduced, the platform stops being for everyone.

Parents as owners, blackmail as a parenting strategy

The most pointed cultural argument in the conversation comes when Zunaira widens the lens beyond product and into family dynamics. “Our South Asian culture — parents feel like they are the owner of the child, which is very unfortunate. You are not the owner of the child. Guide them.” Muzamil extends the thought: an older generation that was not able to pursue what it actually wanted now watches its children go through college and projects its own thwarted ambitions onto them. The civil-engineer parent insists the child become an electrical engineer because, at some point twenty years ago, the electrical engineers were doing better.

Zunaira sharpens it further. The corollary of ownership is blackmail. When the child resists, the emotional pressure begins. When the child eventually fails — because they were pushed into something they were never suited for — the line is always the same: “I told you. Now go find a job. You’re not getting one? I told you so.” She lands it cleanly: the child should be the owner of his failures as well as his successes. Muzamil’s add is that the parent is, in plain terms, “projecting your dreams on an entirely different human being with a different set of experiences.”

The mentorship marketplace Muzamil sketches out

A long stretch in the second half of the conversation is Muzamil thinking out loud about what Merafuture could become. His observation is that the platform already has the hardest thing to build — a top-of-funnel of students who have taken an objective test and now know what kind of professional they want to become. The piece missing, he argues, is a marketplace layer where mentors can plug in, students can browse them by experience level and price, and the platform takes a cut.

He is specific. A free or near-free baseline test could feed three to four hundred thousand top-of-funnel users for every ten thousand who pay for the full test. Within the app, the test result already names a field. Now show the student every mentor in that field, each with their own rate — five hundred rupees an hour at the lower end, five thousand at the higher end. Let the student book, let the platform clip fifteen to twenty percent, and let mentors monetise time they were giving away as free advice anyway.

Zunaira is honest about the constraint. They tried free tests and rolled it back. They have explored bringing industry mentors in, but the unit economics of paid mentorship pull the company up-market, away from the audience she actually wants to serve. Her current compromise is a YouTube section — “Advice from Professionals” — where mentors across fields explain the pros and cons of their work for free, and in-person counselling is reserved for those who can pay for it.

Marketing: kids, parents, and which platform reaches whom

The conversation also gets practical on go-to-market. Zunaira names her marketing problems plainly. Funding is the first. Social media is increasingly pay-to-win — organic reach has collapsed, and a startup’s budget is finite. The second problem is language. Merafuture’s test exists in both Urdu and English because most of Pakistan still consumes education in Urdu, but the marketing question of whether to run campaigns in Urdu, English, or Roman Urdu has no clean answer. The third is persona. The child takes the test; the parent decides whether the child can take the test. Two very different humans, two very different message strategies.

Muzamil’s advice is technical and worth quoting at length. Influencer marketing, he tells her, is for brand affinity, not conversion. “Influencer marketing is like steroids. You pay a ton of cash to an influencer and it’s a hit or miss.” Media buying — proper AB-tested ad campaigns on YouTube, Facebook, Instagram — is the sustainable engine. He maps the platforms to audience segments cleanly: TikTok skews lower- and middle-income; Instagram skews upper-middle and middle with social mobility; YouTube hits the full spectrum, which is exactly why he built Thought Behind Things on it.

He suggests a thirty-second pre-roll on YouTube that simply asks the question the kid is already asking himself: are you confused about your future? Click here, take our test. He does not care what video the kid is about to watch. The ad runs anyway. For Merafuture’s actual customer mix — school owners on LinkedIn, parents on Facebook, students on Instagram — he recommends running different creative for each platform rather than one campaign duplicated everywhere. Zunaira’s daughter, she mentions, is fifteen and on Instagram. Her own peers and older are on Facebook. The school principals come in through LinkedIn. The marketing has to follow the people.

The closing line on scope

By the end of the conversation Zunaira returns to the message she clearly wants parents to hear. Stop chasing scope. “If your child is good at drawing, get them sketchbooks. Get them enrolled into something. If he likes computers, get him into coding. Just guide them. As parents we can only nurture. We cannot do anything else.” Muzamil closes on the same note: it does not really matter if a field seemingly has zero scope. If you are good at it, you could potentially be the wonder kid.

The episode is, in the end, a conversation about replacing a missing institution. Pakistan does not have career counsellors at scale and is not going to have them at scale any time soon. Dr. Zunaira Saqib’s bet is that an objective AI model, trained on people who are already where students want to go, can do the job that no human system in the country is set up to do — provided parents get out of the way long enough to let the test results speak.