AI becomes key lever on company profits, says PwC Cyprus

AI is no longer simply a productivity tool, but is becoming one of the most important sources of operating leverage for companies that know how to use it, according to Yiannis Stavrianos, Senior Manager, Advisory Services at PwC Cyprus. Stavrianos said the shift is already visible among a small group of organisations that have moved beyond experiments and demos, and are now using AI to reshape their cost base, growth model and decision-making processes. “Every CEO has, by now, seen the demos,” he said, pointing to AI tools that summarise documents, draft emails and answer customer questions. However, he explained that the temptation over the past two years has been to treat these tools as “useful additions to the technology stack”, somewhere between “a better search engine and a cleverer assistant”. That view, he said, is “becoming increasingly difficult to defend”. According to Stavrianos, the turning point came in April 2026, when PwC published the results of its global AI Performance Study, based on a survey of 1,217 senior executives across 25 sectors. The study found that “nearly three-quarters of AI’s economic value is already being captured by just 20 per cent of companies”, he said. For Stavrianos, that figure matters because it shows two things at once. AI value is now “real enough to measure”, but the benefits are not being distributed evenly across the market. He said the mistake many companies have made is to view AI through the narrow lens of productivity. “Productivity is a small word,” he said, explaining that it suggests “an extra 5 per cent of output per worker”, similar to what a strong HR programme might also deliver. That framing, he said, has cost organisations “real time and real money” because it encouraged them to place AI on top of unchanged workflows, only to be disappointed when the financial impact proved limited. Stavrianos said the “right framing is different”, adding that “AI, deployed well, is a lever on unit economics”. He explained that AI changes the financial structure of a business in three major ways. It compresses the cost of high-volume customer-facing operations, lowers the marginal cost of growth in knowledge-intensive functions, and improves the economics of decision-heavy functions after losses are taken into account. He said the leaders are “not using fundamentally different models”, but are instead “rebuilding the operating model around them”. Stavrianos pointed first to Klarna, saying the company’s OpenAI-powered customer service agent showed how AI can alter the cost of customer operations. Within months of launching the agent in early 2024, Klarna was handling two-thirds of all customer service chats with AI, a volume the company described as equal to the workload of around 700 full-time agents. Average resolution time fell from 11 minutes to under two, while repeat inquiries dropped by 25 per cent. By the third quarter of 2025, he said, the company had reported around $60 million in cost savings from that single deployment, while revenue rose 26 per cent year on year. Stavrianos described this as operating leverage “in the precise sense the term has meant for a century”. At the same time, he noted that Klarna later moved away from its initial “AI-only” stance and rehired humans for complex cases, after its chief executive acknowledged that the company had “overpivoted” on cost reduction. The savings continued, he said, explaining that the architecture which worked was ultimately hybrid, with humans involved at the right moments. He said the leaders are not “the organisations that automate the most”, but those that “automate the right things”. The second example, he said, is GitHub Copilot, which illustrates how AI can reduce the cost of growth. Stavrianos said GitHub’s controlled research found that developers completed standardised coding tasks up to 55 per cent faster with Copilot. The tool is now used by more than 20 million developers, including around 90 per cent of the Fortune 100, while an average of 46 per cent of the code written by its users is now AI-generated. Stavrianos said this is “no longer an early-adopter phenomenon”, but “the new baseline of engineering productivity in the largest companies in the world”. For chief financial officers, he explained, the message is clear. This is not simply a cost-saving story, but “a revenue-enabling one”, because it creates a structurally lower cost of growth for any technology-led business. The third example, Stavrianos said, is JPMorgan, where AI is increasingly embedded in decision-heavy functions. He referred to comments made by chief executive Jamie Dimon to Bloomberg in October 2025, when Dimon said the bank’s roughly $2 billion annual AI investment was generating about $2 billion of measurable benefit each year. Dimon described this as “the tip of the iceberg”, Stavrianos said. He added that software engineers at the bank were 10 to 20 per cent more productive, operations staff were handling 6 per cent more accounts each, and per-unit fraud handling costs were down by around 11 per cent. At JPMorgan’s scale, he said, these gains compound year after year into “a structural advantage that competitors will struggle to close”. For Stavrianos, the pattern matters more than any single number, because AI shows up “not as a discrete project but as an operating layer”. However, he also warned that most organisations are still far from reaching that point. Citing MIT’s 2025 State of AI in Business report, Stavrianos said an analysis of more than 300 enterprise AI deployments found that around 95 per cent of pilot programmes failed to deliver measurable financial returns. He also referred to S&P Global figures showing that 42 per cent of companies scrapped most of their AI initiatives in 2025, up from 17 per cent the year before. Stavrianos said the numbers are “not contradictory”, but describe “different populations”. The leading companies, he explained, have deployed AI and are measuring its impact. The rest are still attempting pilot projects that never reach production. According to Stavrianos, the gap is not explained by access to better technology, as most serious players are using the same handful of foundation models. Rather, he said, “AI is a business problem dressed up as a technology problem.” The companies that are succeeding, he added, carried out the operating-model work first, while others are still treating AI as a procurement exercise. Stavrianos said AI is now becoming a defining factor in enterprise competitiveness, with the companies that capture it well likely to build advantages in cost, growth and margins over the next two to three years. Those that fail to do so, he warned, risk falling further behind. For most leadership teams, he said, the question is no longer whether AI delivers value, as leading companies have already settled that point. Instead, the more pressing question is why capturing that value has proved so much harder than the demos suggested, and what the winners understand that others do not. The answer, Stavrianos concluded, “is the subject of the next article in this series.”