New Delhi, Aug. 10 -- Enterprises struggling to generate returns from artificial intelligence (AI) need to shift their focus from experimenting with models to solving clearly defined business problems, as rapid advances in AI are making technology-led competitive advantages increasingly short-lived, according to Dinesh Venugopal, CEO of Tenarai, formerly IT services firm Infogain.
The company is shifting its strategy away from building standalone AI platforms towards solving specific business problems, as rapid advances in AI models shorten the competitive advantage companies can derive from technology alone, chief executive Dinesh Venugopal said.
Venugopal said a large number of companies remain caught in what Nvidia CEO Jensen Huang has described as "pilot purgatory", with AI projects failing to progress from proof-of-concept deployments to production environments.
"Most people start with the wrong question. They ask, 'can AI help us?' when they should ask, 'what business outcome can AI help us achieve?'" Venugopal said.
He cited a recent survey indicating that around 55% of CXOs are not seeing tangible value or return on investment from AI. According to him, the challenge is increasingly less about the capability of AI models and more about data readiness, governance, security, integration and organisational change required to deploy them at scale.
Technology advantage is getting shorter
The rapid pace of model development is also reducing the shelf life of technology differentiation.
Venugopal said Tenarai built an AI platform in July last year that initially drew strong customer interest, but competing offerings had caught up within about eight weeks as newer models became available.
"The technology advantage had a shelf life of about eight weeks," he said, adding that the experience prompted the company to shift its strategy from primarily building platforms to solving specific business problems.
This could have wider implications for India's IT services industry, which has traditionally relied on building large pools of employees trained around specific technology stacks.
"The knowledge-worker era is over," Venugopal said. "You can't train your way through AI." Instead, he expects demand to move towards professionals capable of combining technology expertise with industry and customer context to deliver measurable outcomes.
Tenarai creates new AI roles
Tenarai has introduced roles such as Forward Deployed Engineers (FDEs) and Agentic Design Leads as part of this shift. The company opened applications for its FDE programme to more than 6,000 employees and is cross-training workers for these roles as part of its longer-term talent strategy.
Venugopal said India's AI talent challenge is therefore not primarily a shortage of engineers or people familiar with models.
"The gap isn't engineering headcount or model knowledge; both are abundant. People who can solve a problem and deliver business outcomes are in demand," he said.
The emergence of such roles could also accelerate India's transition from being primarily a technology delivery base to becoming a centre for AI engineering and decision-making for global companies, he added.
FDEs, for instance, typically work closely with customers to combine business context, industry expertise and technology capabilities while taking responsibility for solutions deployed in production.
"That combination of context and decision-making is becoming distributed. It goes where the talent is. India can absolutely become that core," Venugopal said.
From chatbots to decision-led AI
Enterprise AI adoption has also changed significantly from the chatbot-heavy experimentation of 2023 and 2024, according to Venugopal.
He cited the example of a retailer that initially approached Tenarai seeking a chatbot. Discussions subsequently established that the underlying objective was to reduce customer call waiting times. Tenarai and a partner then developed a broader solution incorporating customer intelligence and enterprise context rather than treating the chatbot itself as the end product.
The shift is also changing how companies should measure AI investments. Metrics such as the number of pilots, deployment timelines and user feedback provide limited indication of whether AI is creating financial value, Venugopal said.
Boards are increasingly looking instead at measures such as new or improved revenue streams, margins, customer acquisition, retention and satisfaction.
"The lens that matters is proof of value in production and outcomes to the business," he said.
For enterprises seeking to move AI beyond experimentation, Venugopal said the approach should begin with defining the business outcome, followed by understanding the organisation's operating and industry context before choosing the technology.
"Start with outcomes, go back and understand the customer's context, understand the domain they operate in, and only then design the solution," he said.
Published by HT Digital Content Services with permission from TechCircle.