According to the Commission on Higher Education, approximately 805,000 students graduated from university in 2023-2024. Some have spent the past year looking for jobs that don’t exist. So that was the question I kept asking when I spoke to Chief Technology Officer Vaibhav Vora at Ascendion’s Makati office this week. Ascendion describes itself as an AI-native, platform-driven engineering services company where AI is transforming offshore operations that have employed Filipinos for 20 years. Is it a better job or no job for someone just starting out?
Vora’s case begins at the gap. From 2022 onwards, corporate technology spending will increase by about 8% per year, but productivity growth will be only 2%. These are McKinsey’s U.S. numbers, and McKinsey itself admits the relationship is not accurate. When asked if this pattern holds true here, Vora said it extends beyond the United States to Europe and the Asia-Pacific region, including the Philippines. He explains that most companies keep AI confined to pilot projects rather than implementing it across the business. He says Ascendion’s job is to move things from pilot to production. He calls the platform AAVA. This is an in-house system that runs AI agents throughout the software development cycle rather than one task at a time.
Since Ascendion sells repairs for older systems, I asked if that was a diagnostic or a sales pitch. Modernization is as old as the industry, he said, and the difference lies in how it’s done. He pointed to the overhaul of U.S. banks that Ascendion’s AI agents enabled at a third of the cost.
Ascendion has been operating in the Philippines since 2013 and has worked with more than 80 local clients, including global capability centers (GCCs) operated by multinational companies in the Philippines. When asked about employment, Vora answered candidly. “For us, it’s not a job replacement,” he said. His clearest example is the customer service operations the company runs for multinational corporations in the Philippines. Before AI, the average call took 20 minutes. Using AI tools, he said, that time was reduced to two to three minutes, and the number of employees was not reduced. Agents can now see a caller’s history and preferences on one screen, instead of pulling information from 10 separate systems. He says entry-level roles aren’t going away. In his words, it undergoes “remodeling.” Today’s hires need agile, contextual engineering skills and data science capabilities that were not required for this job 25 years ago.
That’s the hopeful version. More difficult came from Dr. Paul Roehrig, Ascendion’s chief strategy and marketing officer. Humans are smart enough to understand that the same task may require less effort, he said. “And that may be true.”
What complicates reskilling is where the demand will come from. Vora said he had met that morning with the head of GCC, which was under pressure from corporate headquarters to cut costs. He said AI is the mechanism to achieve these savings while freeing up people for more impactful work. Ascendion’s fact sheet puts it in simpler terms, pointing to a project that has made work “50 percent more economical” and “freed up” 6,000 engineers. Retraining and cost cutting are both realities, and workers are caught in the middle.
This is not an abstract problem. Two weeks ago, I wrote about how hiring new employees in business process outsourcing is becoming tougher as AI absorbs the menial tasks that once trained new graduates. We previously told you about a Filipino freelance writer whose income decreased as his clients transitioned to AI. Vora is right that people can be retrained. However, there are very few entry-level jobs where you can learn the job in the first place.
So what should graduate students study? Vora, an engineer whose 18-year-old son is about to go to college, gives the same answer he would give his own child. Build a strong foundation for analytical reasoning. Next, move beyond your old programming language to Python and build your foundation in data science. Become proficient with tools such as Claude and Microsoft Copilot. He built a program he called “10X” around early career recruiting. Nine out of 10 members of the team are entry-level employees, he said, and the UK’s highest-performing engineers only graduated last year. He believes new graduates can adopt the tools the fastest.
Roerig said the fear is real. In the United States, commencement speakers are booed when they mention AI. His answer is that the work that will be done, such as legacy systems that organizations don’t have enough people to modernize, is greater than the work that will be automated away. He might be right. But he’s betting on a job that doesn’t exist yet, and his company sells the tools to create it.
What I can say to graduates is even simpler. The stage they are trying to reach is even higher. Vora’s advice to his son is the closest thing to a map. Learn the basics, then learn how to ask machines the right questions. His son plans to enter university with the map in hand. How many of our 805,000 graduates will reach that level? What about those who couldn’t be reached? Their answers never fully resolved the question.
