Errors and high costs are some of the reasons why GenAI does not advance beyond the conceptual stage

Leading industry executives and analysts say a high error rate, high implementation cost and the relative newness of the technology are the main reasons for weak adoption of GenAI by enterprises.

“First, the accuracy of the results is of critical importance. And second, the adoption of that solution is of critical importance. And adoption only happens if the accuracy levels are high enough to give you a clear competitive advantage,” said Rohit Kapoor, CEO of ExlService Holdings Inc., in his post-earnings interaction with analysts in May 2024.

New York-based ExlService is a Nasdaq-listed BPO company that closed the quarter ended June 2024 with $448 million in revenue, up 2.7% from the previous three months. Its revenue for the year ended December 2023 was $1.6 billion, up 15.5% from 2022.

Kapoor, who is among the first CEOs to highlight the high error rate in GenAI-led solutions, said the goal was to improve their accuracy.

“So when we start with a new solution, the accuracy levels are very low. They are 65% to 70%, but we need to increase them to 90% or 95%, and that comes with deep domain knowledge and understanding of how and what the data is telling us, and therefore fine-tuning our algorithms and our ability to integrate them into the workflow in a way that makes them really useful for every person who adopts these solutions,” Kapoor said.

GenAI caused a stir after the launch of ChatGPT in November 2022 catapulted the new technology into the boardroom discussions of the world’s largest companies. It seemed like its widespread adoption was a foregone conclusion.

Nearly two years later, GenAI has failed to live up to expectations.

In an interview with Mint On July 12, K. Krithivasan, CEO of India’s largest IT services company Tata Consultancy Services Ltd (TCS), said that offering customers differentiation is key, rather than following the trend. However, the path to adoption remains uncertain.

GenAI in outsourcing companies

Experts have questioned the adoption of GenAI by outsourcing companies as most of the tests carried out by these companies have not reached the production phase.

“According to industry feedback, over 85% of proofs of concept around generative AI have failed to move into production. We believe this contributes to a disparity between generative AI investment in hardware and training versus actual revenue generation by IT service providers and software application vendors,” reads a note from BMO Capital Markets analyst Keith Bachman dated July 31.

The shift towards artificial intelligence in IT companies

Most IT services companies in their latest earnings commentaries have said that they have trained their employees in AI and GenAI, without going into depth on those training modules.

Gartner has said that nearly a third of GenAI projects would not see the light of day beyond the proof-of-concept stage.

“At least 30% of generative AI (GenAI) projects will be abandoned after proof of concept by the end of 2025, due to poor data quality, inadequate risk controls, rising costs or unclear business value,” the IT research and consulting firm said on July 29.

While accuracy is a concern, stakeholders are wary of the new technology’s novelty. Analysts said businesses were unsure of GenAI’s actual adoption due to its relative newness.

“Only 5% of GenAI has moved into production at large enterprises, but this number will increase significantly as real business cases emerge. Remember, we’re only 18 months into this and enterprises are still figuring out where it adds the most impact and value,” said Phil Fersht, CEO of HFS Research, a US-based outsourcing research firm.

Additionally, experts said the cost of implementation was too high for GenAI adoption to be immediate.

“To be commercialized, GenAI proofs of concept need to be integrated into the IT infrastructure of outsourcing companies. This requires multiple building blocks such as accurate data layering. In addition, there is an implementation cost that will force outsourcing companies to spend significant expenditure on this integration. Adoption will take a couple of years,” said an industry expert working at a BPO firm on condition of anonymity.

Mint On July 4, citing analysts, it was reported that GenAI posed a threat to BPO units as the new technology, which can create content in audio, visual or written format through a message, can replace routine human work. BPOs offer customer support to their clients, much like how a call center works.

In fact, global IT services companies including Accenture Plc and Cognizant Technology Solutions Corp. have highlighted GenAI as a risk in their annual reports.

“We are increasingly using AI-based technologies, including GenAI, in our offerings to clients and in our own internal operations. As with many innovations, AI presents risks and challenges that could negatively impact our business. The development, adoption and use of AI technologies is still in its early stages and ineffective or inappropriate AI development or deployment practices by us, our clients or third parties with whom we do business could have unintended consequences,” Cognizant said in its annual report for the year ended December 2023.

Cognizant follows a January-December financial calendar, while Indian IT services companies follow an April-March financial calendar.

Caution with GenAI

Wariness toward GenAI was no different at home.

“GenAI, large language models, and Metaverse could lead to legal liabilities for plagiarism, deepfakes, and privacy and copyright infringement issues. The effectiveness of AI models depends on the quality of the data they are trained on. Risks of accuracy, bias, or impartiality could lead to reputational damage and legal liabilities. GenAI technologies could disrupt software development and testing activities, changing customer expectations in the near term,” reads TCS’s annual report for the year ended March 31.

TCS’s Krithivasan said customers are taking a calibrated approach in mapping out their GenAI adoption roadmap.

“While interest is strong, organizations are taking a calibrated approach to measuring risk potential and organizational impact as they chart their roadmap for GenAI adoption,” Krithivasan said during the company’s post-earnings interaction with analysts on July 11.

However, despite the current slow pace of GenAI adoption, experts said the technology will likely replace basic human functions in customer experience areas.

“The work that will be affected will mainly focus on areas such as customer experience, marketing and IT development. The current workforce will be reduced by 20-50%, and those remaining will be trained to use the tools,” Fersht said.

The BPO industry expert quoted above said that humans will always be part of the cycle to deal with complex queries even as chatbots replace basic customer experience functions.

“What companies have started to do is integrate GenAI into a human-centric process. It is now a process where humans and AI are combined, which improves the efficiency of the person involved and the process. We are at a stage where it will never be a binary movement. Humans will always be involved,” the expert added.