As artificial intelligence reshapes software development, creative work and business operations, Pakistan’s technology industry is entering a significant transition. We spoke with Abdullah Rashid, a Pakistan-based technology entrepreneur and software engineer, about how AI is changing the country’s services economy, what makes an AI product genuinely useful, and where he sees the next wave of opportunity.
Abdullah Rashid is the co-founder and Chief Technology Officer of ClearTraced.com, an AI Swedish climate and ESG technology company developing AI-driven systems for collecting, analyzing, verifying and tracing sustainability information. The company has worked with organizations including Bonnier, Dagens Industri, Millistream and Global Child Forum. Abdullah has also founded and led multiple technology ventures, including Fluently AI, a voice-based AI learning platform that was later acquired, and Metasense.dev, an AI services agency he runs that has worked with more than 30 clients globally. His work has spanned generative AI, voice AI, computer vision, SaaS and mobile products, while his online courses have reached more than 60,000 students. More recently, he has begun conducting independent research into AI safety, interpretability and the reliability of AI systems.
With AI rapidly changing the economics of software development and technology services, we spoke with Abdullah about what the shift could mean for Pakistan and where he believes the next opportunities will emerge.
Pakistan has historically relied heavily on software services and freelancing. Does AI threaten that model?
I think there is a tendency to look at the first-order effect of AI and assume that it is the whole story.
Pakistan has built a significant technology-services and freelancing economy, and AI is already changing that model. Entry-level development work is becoming increasingly automated, while developers who use AI effectively can become significantly more productive. At the same time, some of the traditional project inflow is declining.
It is easy to look at that and conclude that AI is going to finish the industry.
I actually think the opposite may happen.
What is changing is the kind of work the industry can provide. We are likely to see a new layer of AI-enabled services emerge: AI-assisted development, AI creative production, generation of commercial assets, workflow automation, data processing and other services that were either too expensive or too slow to deliver previously.
So while the transition may be disruptive, particularly for entry-level roles, disruption does not necessarily mean the disappearance of the industry. It can mean that the industry evolves into something more sophisticated.
For Pakistan, the opportunity is to move from being primarily a provider of human labor toward becoming a provider of AI-enhanced outcomes.
There is a lot of excitement around AI agents and increasingly capable models. What makes an AI product genuinely useful?
The most important question is not what an AI system can theoretically do. It is what it actually changes for the person using it.
It is very easy to say that an agent can perform ten different tasks or that a model can automate an entire workflow. But the more meaningful measurement is much simpler: how much did it improve someone’s productivity?
Sometimes the most valuable use of AI is not the most impressive demonstration. It is the unglamorous work.
Data preparation, repetitive analysis, administrative tasks, content variations, internal workflows and other forms of routine work can consume enormous amounts of time. AI can be extremely useful in reducing that “grunt work” and allowing people to spend more time on decisions and tasks that require judgment.
I think we will increasingly move away from measuring AI primarily by how flashy the demo is and toward measuring it by how much useful work it enables people to accomplish.
An AI system does not have to replace an entire employee or workflow to create significant value. Sometimes making a person 20 or 30 percent more productive at an important task is already a major business outcome.
You have increasingly become interested in AI research. Why is it important for people in the industry to contribute to the field rather than simply consume AI?
AI is still a relatively new field, and that creates an unusual opportunity for people working with the technology today to contribute to understanding it.
There is a tendency for countries and companies to position themselves primarily as consumers of technology developed elsewhere. I think AI presents an opportunity to approach that differently.
We should not only be asking how we can use the newest models, APIs and tools. We should also be asking what we can add to the advancement of the field.
That can mean research, benchmarks, evaluations, engineering, safety work, new applications or simply documenting problems that others have not yet understood properly.
For me, that is part of the motivation behind becoming more involved in AI safety and interpretability research. Building products gives you practical exposure to what these systems can and cannot do. Research can help turn those observations into something more general and useful to the wider field.
Pakistan does not need to wait for AI to become a mature industry before contributing to it.
There is an opportunity right now to build locally, publish internationally and contribute evidence, tools and ideas to a field that is still being defined.
What do you think the next phase of AI adoption will actually look like?
I think the near-term opportunity is less about replacing entire systems and more about understanding how AI can be integrated safely into the systems we already have.
There is still a significant human-in-the-loop overhead because organizations are adopting the technology while also learning where it can fail, where oversight is necessary and what kinds of decisions should remain with people.
That creates an important transitionary period.
Companies have to figure out how AI fits into existing software, business processes and human workflows without assuming that automation is automatically reliable. There are technical, organizational and safety questions that come with that integration.
A lot of opportunity exists in that layer of transition: building the infrastructure, tools and processes that allow organizations to adopt AI responsibly and productively.
Whether AI eventually takes over much larger portions of work is still an uncertain question. I don’t think anyone can make a particularly confident prediction about the full extent of that transition yet.
What is much clearer is that we are already in the period where people and organizations have to figure out how to work with AI effectively.
And that period itself is going to create an enormous amount of opportunity.
The bigger picture
Pakistan’s AI opportunity may ultimately have less to do with producing the next headline-grabbing model and more to do with what the country does with the technology.
The country already has a large technology-services workforce, a growing startup ecosystem and a substantial base of software engineers and freelancers. AI will put pressure on parts of that existing economy, particularly work that depends primarily on inexpensive human execution.
But the same technology can also allow that workforce to move further up the value chain.
The companies that benefit most may not necessarily be the ones with the most impressive AI demos. They may be the ones that understand a specific problem, integrate AI into an existing workflow and produce a measurable improvement in how people work.
For Abdullah, that transition represents both a challenge and an opportunity.
The question is no longer simply whether Pakistan can participate in the AI economy.
It is whether its engineers, founders and companies can move quickly enough to help shape what that economy becomes.
