Traditional AI deployment of 70% companies has failed: BCG X executives

AI and GAI are beyond providing just efficiencies in data management, etc. It is about the transformation of business operations, said Sylvain Duranton, MD and senior partner, BCGX

Sylvain Duranton, managing director & senior partner, and global leader, BCG X, and Nipun Kalra, managing director & partner, and head of BCG X India, in an interview with Shivani Shinde at BCG‘s Mumbai office say the implementation of generative AI within enterprises is not just about tech and tools but 70 per cent of it is about people too. They talk about how India is at an advantageous position when it comes to generative AI. Edited excerpts:

Where is India in the generative AI (GAI) cycle?
 

We did a survey recently, globally and in India, on AI and all other technologies, and found that India stood out. India is among the top three countries where people are optimistic about AI and GAI. India is among the bottom three when it comes to concerns and challenges regarding deployment of AI and GAI. I think this will be an asset for India, which can have a good run.

Until now tech discussion was around cloud. Now the focus has shifted to GAI. How should business look at the tech landscape?

Have companies started to move to cloud? Yes. Have they got the benefits that were promised to them? No. I think there has been a good migration to cloud and companies should focus on this.That said, AI and GAI are beyond providing just efficiencies in data management, etc. It is about the transformation of business operations. Hence deploying AI is challenging.

But it will take longer than what most companies think. Many people think Chat GPT is very easy to deploy, and there is hardly any engineering. That is a myth. If you want to deploy this at scale in companies, you need full engineering muscles.

What are some of the challenges in deploying GAI?

Almost 70 per cent of the companies have failed when we look at traditional AI deployment. We are saying AI is not a magic button, it has tech challenges … companies have to build new tech and also bring it into the legacy systems. GAI will be the same, with twice the impact and potential but twice more headaches. My advice to companies and chief executive officers is that they need to pick a few areas that will truly disrupt. If companies want to deploy this in everything, they will be lost. They need to pick the process based on their impact on business.

Kalra: From a traditional AI perspective, yes, 70 per cent have not made money. AI-led cases have been piecemeal. GAI is very data-hungry and that is a journey. We met 45 CEOs in India recently, and one refrain is how the employee base can be productive. They are looking at small cases.

Generative AI impacting jobs is one thing, but what does using it mean for employees?

Duranton: Getting the tech right or data correct is just 20 per cent of GAI; 70 per cent is about how people will be adopting this. If you are telling a salesperson the person’s job is no more about fixing a price for a product to sell at but to sell at what a machine has set and convince the client, then it’s a different job. It is not that the job has gone; it has to be done differently. Many people think GAI is only about buying tech and deploying it. No, it’s about transforming what people do.

Worldwide 36 per cent of workers think that AI will take their jobs. That number in India is 50 per cent... I do not believe those numbers are correct. There is a massive worry in the workplace, which needs to be addressed. When we deploy GAI, we need to tell people that they are part of the solution.

Does it mean that we will see more success with GAI?

Kalra: We know that 30 per cent of companies succeed with traditional AI. A maximum of 30 per cent will also succeed with GAI. The difference is that the benefit they will get over their competition will be much bigger.

We know that worldwide 36 per cent of workers think that AI will take their job. That number in India is 50 per cent because AI is much more pervasive in India than in many other countries. I do not believe those numbers are correct. There is a massive worry in the workplace that needs to be addressed. When we deploy GAI we need to tell people that they are part of the solutions. This needs to be addressed, there is a massive gap.