- AI success depends on strong cloud, data and cybersecurity skills as much as generative AI, with 88% of leaders saying workforce upskilling is essential.
- Safe, hands-on experimentation and human oversight are key to integrating AI effectively, shifting work toward augmentation rather than merely automation.
- Real-world projects and practical training are critical, as companies seek to build skills that directly support digital transformation goals.
As enterprises accelerate artificial intelligence adoption, Coursera CTO Mustafa Furniturewala believes success will depend as much on cloud foundations and human skills as on generative AI itself.
In an interview with The AI Innovator, he said that the excitement around generative AI must be accompanied by rock-solid skills to bring out the potential of AI in the enterprise. Otherwise, it is all hype.
In a joint survey with AWS, Coursera found that executives ranked cloud, data, and cybersecurity skills before AI in importance for the next three years. Leaders recognized that talent development is essential to achieving their digital transformation goals.
That makes sense to Furniturewala, who argued that one cannot operationalize AI at scale without robust cloud infrastructure. And when deploying or using AI runs into a problem, as it will, subject matter expertise will be needed to solve the issue, he said.
“Yes, generative AI is an important skill to learn, but the domain skills are as important, if not more important, to be able to use the technology in the right way,” he said.
“We are also not only upskilling on generative AI, but we don’t forget the data skills, the cloud skills, the cybersecurity skills, and we’ll continue upskilling on those as well,” Furniturewala added.
Furniturewala said the survey highlighted this dual challenge: AI creates pressure to move faster, but also raises the bar for continuous learning.
“The expectations of individuals at a company are higher because they have these tools at their disposal, and at the same time, they are going through a learning phase,” he said. “That creates an interesting dynamic where it becomes really important in the short- to medium-term to make sure we are learning how to use this technology well.”
Cloud skills are critically important
In the report, 95% of the over 750 executives surveyed in six countries identified cloud transformation as a key business goal. Six out of 10 ranked cloud skills – such as cloud development and cloud engineering – as more critical than AI skills (47%) for the next three years. Data skills came in at 58% and cybersecurity skills at 54%.
Over half of technology leaders believe that 30% to 50% of their tasks and their teams’ tasks will be automated within three years. Nearly all (99%) expect their developers to use AI to help them code.
Crucially, 88% of business leaders said their AI investments will not succeed without accompanying skills development of their workforce. Nearly eight out of 10 said upskilling existing employees was “essential” to meeting their transformation goals in the next 12 to 18 months.
But the training can’t just be theoretical. Sixty percent of respondents said real-world projects that are directly relevant to employees’ work are the “most valuable element” in technical training. Rounding out the top three are practical skills assessment and risk free experimentation.
Furniturewala agreed: “Learning by building and practical, hands-on assessments will become more important.”
Coursera’s own engineering teams have been an early test case. Adoption of AI tools rose from 30% to 100% in short order. And it was voluntary.
“I actually didn’t mandate my team to use these tools,” Furniturewala said. “I instead enabled a culture of adoption and safe experimentation, and that actually led to 100% adoption faster.” Workers’ natural curiosity and productivity gains that were “impossible to ignore” led the way.
But he stressed that experimentation needs boundaries. “The safe experimentation with this technology is extremely important,” Furniturewala said. At Coursera, “we have a lot of content on how to use this technology in a safe way, by using evals and by using domain experts, which brings us back to the point of upskilling on those skills.”
He cautioned against rushing full workflows into AI systems. “Just making sure that you’re not just leaving AI to automate really long workflows because of the challenges around hallucination and (other risks). So doing that safe experimentation, again, is critical.”
Despite concerns about automation, Furniturewala said AI is not eliminating work but shifting its nature. “It’s augmentation, not automation. So the human oversight is still very important.
Fantastic read! 👏 I really appreciate how clearly you explained the topic—your writing not only shows expertise but also makes the subject approachable for a wide audience. It’s rare to come across content that feels both insightful and practical at the same time. At explodingbrands.de we run a growing directory site in Germany that features businesses from many different categories. That’s why I truly value articles like yours, because they highlight how knowledge and visibility can create stronger connections between people, services, and opportunities.Keep up the great work—I’ll definitely be checking back for more of your insights! 🚀
Thank you! Very kind of you to say so.