GenAI is a long-term play, not an instant gratification project

We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.

– Roy Amara, American researcher, scientist, and futurist

The biggest buzzword today is “GenAI” (generative artificial intelligence). While it has seen significant adoption, we are still at the start of a journey. GenAI promises to transform work, upend operations as we know them, and unleash a productivity boost, but many challenges remain. Data privacy, technical challenges, talent shortages, and ethical challenges are just some of the concerns that CEOs must navigate.

CEO Nitin Rakesh is navigating the journey for Mphasis, a global software and technology solutions company. We had the chance to sit down with him to discuss how the industry is navigating the challenge that GenAI poses, the need for upskilling the workforce, and how the Indian IT ecosystem stacks up to respond to the challenges ahead.

Some of the key highlights from the conversation follow.

GenAI is a double-edged sword that can unleash a productivity shift

While GenAI has taken the world by storm, it didn’t arrive suddenly. The years leading up to GenAI had seen growing digitization, cloud services adoption, improvements in artificial intelligence, and more. The congruence of several developments, including increasing compute power, the availability of large data sets, and the ability for enterprises to invest large resources into solving problems, resulted in GenAI.

Like all disruptive tech, GenAI is something of a double-edged sword. It has the potential to reshape, rewire, and reset how business is done—an exciting prospect for enterprises. However, those who aren’t acting urgently on the issue are in danger of being left behind.

Being left behind in the GenAI revolution could be catastrophic. GenAI has the potential to create a significant productivity boost and transform the future of work on a scale similar to that of the advent of the internet. Enterprises hoping to capitalize on GenAI’s benefits will need to invest resources, but it can reshape how work is done today. Almost every aspect of the value chain has the potential to adopt AI, impacting how jobs are done and the skillsets required for them, reducing the need for manual interventions, and even changing how software engineers work.

GenAI is a long-term play; don’t look for instant results

The industry shouldn’t be looking at GenAI as a 12-to-18-month initiative. It should see the coming year as the start of a transformative journey. It would be wise to keep in mind Roy Amara’s law, which states that humans tend to overestimate the impact of technology in the short run and underestimate it in the long run.

It will take time for enterprises to set up data pipelines, ensure clean data availability, install model management platforms, and achieve these objectives in a traceable, secure, and ethical manner. In many ways, the next year is a pivotal period for enterprises to build a solid foundation to serve as the basis for transformation. Enterprises will likely embark on one or two large enterprise application areas; fields such as customer experience, support, and the help desk are all forerunners for reinventing operations.

Privacy is among the top boardroom concerns

Upon GenAI’s launch, the first instinct for many boardrooms was to express concerns about data security, which continues to be a key focus area. Many industries, such as banking, healthcare, and insurance, are highly regulated and need robust data security mechanisms. Over the next six to 12 months, the focus will be on ensuring that the right infrastructure, pipelines, and relationships with providers are in place.

It is an undeniable fact that GenAI is now a boardroom conversation topic. Almost every enterprise has a point of view being driven from the top down. The consensus among enterprises is to find a few large impact areas instead of executing small pilots and proofs of concept.

Public or industry LLMs? The answer lies in the middle

The debate about whether to use a public LLM or start thinking about an industry-specific LLM is likely to continue for a while. Public LLMs are trained on a large volume of public data and generate answers to almost any question; however, the data they are trained on may not always be accurate. Private LLMs are trained on industry-specific data that isn’t available to the public, and as a result, they provide greater industry depth in responses. An example of an industry-specific LLM is Bloomberg GPT, which is trained on finance-specific information from the company’s repository. The optimal solution will likely lie between public and industry case LLMs, with some applications requiring industry-specific LLMs while others will work on public LLMs.

AI will not replace humans; it will replace humans who don’t use AI

Addressing the ongoing debate about whether or not AI will replace humans, Nitin believes that AI will not replace humans; however, humans who don’t upskill and learn to use AI could be replaced. As a result, reskilling the existing workforce to take advantage of AI will be a key focus area for enterprises.

Earlier this year, Mphasis launched Mphasis.ai, a business unit with a charter of embedding AI in every archetype it runs, including customer experience transformation, DevOps, application development, and developer productivity.

India can be the world’s largest AI workforce

While disruption poses a threat to India’s established IT industry, it is also one of the best-placed countries to respond to the AI revolution due to its strong local economy and ability to create a large pool of talent in almost any industry. The IT industry has successfully embedded itself into India’s educational system from an engineering and design standpoint. As long as there is clarity about the large application areas and solutions align with skills, India could boast one of the world’s largest AI workforces within five years.

The Bottom Line: Cut through the market noise and focus on priority areas.

GenAI could transform how enterprises work, but with its potential comes a lot of noise. Companies will need to identify the key areas they would like to start with and invest resources accordingly. Implementing GenAI should be seen as a long-term play, not a 12-to-18-month project, to provide time to mitigate concerns regarding privacy, data security, and other issues. As there are frequent new developments in the field, companies will need to be nimble and quick in responding to developments. Now is not the time to take one’s eye off the prize.