From AI Pilots to Production: Kyndryl Sees Consulting Demand Accelerate
"We're getting demand from everyone, starting from SMBs all the way through large enterprises," says Dr. Vishnu Nanduri, Head of AI & Innovation at Kyndryl.
For Dr. Vishnu Nanduri, Head of AI & Innovation at Kyndryl, the IT services provider continues to see more consulting requests from customers, especially in helping them build a strategy around AI. This includes everything from agentic to generative AI in terms of implementing the strategy, developing use cases as well as understanding the different solutions available.
“We continue to help customers design, implement, manage, and run AI for them. There are clients we help with the consulting phase, build the solution for them, and also manage the solution for them. And we're getting demand from everyone, starting from SMBs all the way through large enterprises. While the kind of questions and the kind of use cases are very different based on the industry, the demand is very strong for getting AI services,” Dr Nanduri explained in an interview with CRN Asia.
[Read more: Kyndryl: APAC enterprises must close modernization gap as agentic AI accelerates]
Dr Nanduri also mentioned that a lot of clients are also upskilling rapidly. Unlike the past, he shared that almost everyone in an organization understands the dynamics of technology. This is also why Kyndryl is seeing more demand for consulting as companies want to move beyond prototypes to production and also have sufficient skills for it.
At the same time, he also highlighted that sovereign AI is starting to feature in conversations with customers.
“We do have consulting services and a solution offering around sovereign AI and sovereign cloud services. We are using the technology that is available in the market with the cloud provider that's available in the market to have those conversations. Many of them are coming out with sovereign solutions in different countries to make sure that the data stays there. Nothing goes out. So, I think we as a consulting company, we are providing those services and we're happy to speak to clients about it,” he added.
Defining real AI use cases
While all this sounds positive, Dr Nanduri felt that there is a also trend of organizations using AI for use cases that don’t really need AI.
“Don't use AI everywhere, just use it where it's required. I do see an overuse of AI where it really is not needed, especially for very simple things. For example, people should really write their own emails. So, I think there is definitely overuse in some cases. Because if you just leave out the productivity angle aside for a second, it is really just sort of step up to more business function kind of use cases,” he said.
He added that there are some cases where the business outcome is not thought of before the company gets started. These companies are doing AI because they just want to be part of the bandwagon as opposed to getting the business outcome first and working backwards to the model.
“What we advise in Kyndryl is that companies need to start with the business outcome first and move backwards. We then figure out what model to use. Is it an AI? Is it not AI? If it's AI, is it a traditional AI, more machine learning? Is it computer vision or more? I don’t see these kinds of mistakes done by the very large enterprises but maybe the small and the medium ones,” he said.
When asked how Kyndryl is also using AI for its operations, Dr Nanduri shared that the company is using AI very cautiously, especially with the agentic coding tools that they have at their disposal.
“So, the first thing we do is set up a proper harness to make sure that any of the code that is developed is properly vetted downstream by a human reviewer. There's nothing that moves forward to build applications without it being reviewed, checked, processed and secured. That is a harness that we have set up within the team. We are starting to leverage these coding tools over the last few months, and we have set up a harness properly. But we are using it very cautiously to make sure that it's not just written and deployed automatically. There is a proper process that we have,” he explained.
Specifically, Kyndryl has a 12-step process that it has defined for the use of AI within the organization, as well as within teams in the company.