A couple of weeks ago Accenture CEO Julie Sweet doubled down on the services giant’s commitment to generative AI initiatives, both internally and externally, citing the technology as a major future revenue stream. Inevitably it’s the same story at the leading Indian outsourcing and services companies.
Earlier this month, Tata Consultancy Services (TCS) expanded its links with Azure OpenAI, the venture between Microsoft and OpenAI. It also staked a claim to be one of the largest providers of AI-trainer workforce globally, with over 100,000 employees trained to date.
With more than 250 generative AI opportunities in the pipeline, CEO K. Krithivasan pitches the tech as today's main executive talking point:
It continues to dominate our conversations with IT and business leaders in every market. There is tremendous interest in harnessing its power to drive productivity and enhance customer experience. We are co-innovating with clients across multiple industry verticals, executing proofs.
He cites client engagements exploring the use of generative AI for contract administration and customer service automation, as well as task automation around knowledge abstraction and content creation:
We are now seeing a progressive increase in the complexity and sophistication of gen AI use cases, from simple knowledge discovery use cases and chatbots to complex ones, such as augmentation solutions for financial advisors and wealth management strategists, automated underwriting for insurance policies, AI-led molecule discovery, as well as engineering design space explorations for automotive and gas turbine.
Slightly more mature clients are taking a broader enterprise-wide approach to generative AI deployment, he adds:
For a European airline, TCS is helping re-imagine its core business functions, leveraging gen AI and analytics to create a future-ready enterprise. The TCS team co-presented with the CTO to the Board and is now engaged in helping draw up a holistic gen AI strategy. A leading US-based specialty retailer is partnering with TCS for multiple gen AI-led interventions to enrich customers as well as employee interactions, including customer request handling, virtual assistant for shopping, knowledge management, designs for creatives, marketing and sales collaterals, and code enhancements.
Some other organizations are using generative AI to re-imagine an entire activity. A good example is a client in the construction industry for whom we are using gen AI to prepare the preliminary architectural plans for a new building project. The AI-generated designs are then validated and detailed out by the human team. We see similar opportunities in product innovation where generative AI can visualize multiple new form factors for a product in the manufacturing industry or perhaps new types of packaging in the CPG industry, significantly increasing and accelerating the design process.
And while these are examples of standalone deployment of generative AI, Krithivasan says TCS is also beginning to see larger transformational projects that involve multiple. technologies, where generative AI is “infused” to address specific parts of the value chain:
We believe that as the technology matures, we will see more instances of activity-level automation and eventually end-to-end value chain transformation using generative AI in combination with other technologies. That is when the mainstream adoption of this technology will really take off and the full potential of this technology will be realized.
At Wipro, CEO Thierry Delaporte points to the firm having trained 180,000 employees in basic generative AI principles in just one quarter:
We're rapidly integrating gen AI into our processes, our solutions and our offerings, thousands of our employees have or are starting to use gen AI. In the HR functions, our teams are seeing significant productivity gains by using gen AI for candidate background verification. And in marketing, we are using gen AI for content generation and translation, tasks that used to take hours earlier now takes a minutes. In sales, we're deploying gen AI for research to improve sales collaboration and to generate RFI responses.
[In engineering], gen AI is helping with software development and lifecycle automation. One of the areas with the biggest productivity gain is in quality engineering and quality assurance testing. Gen AI is helping with scenario creation, cogeneration, synthetic data creation as well as execution at scale. Initial pilots of these Gen AI apps have been so successful that we are now rolling them out to all our employees.
As for client-facing activities:
Gen AI is now a part of every client conversation, there is tremendous interest in exploring new use cases as well as understanding the benefits and implications of this technology. Today, we are seeing a doubling of gen AI active projects than we did just one quarter ago. For now, we're seeing rapid adoption in healthcare, consumer and financial services, but also in high tech and utilities.
He highlights a couple of examples:
We are working with a US-based health insurer to deploy a gen AI-based chatbot for their agents. We are developing a solution that is fine-tuned to be more contextual so that agents can provide more personalized assistance to every member. This solution is driving 30% to 40% reduction in operation costs, significant improvements in agent productivity and improving Net Promoter Scores. Another example, we're working with a European multinational telecom company to unlock value from data. Working with different vendor tools and software kits, we are generating high-quality synthetic data, which allows the client to not only increase cross border collaboration, but also mitigate buyers and eliminate distribution limitations that exist in real data.
And this is only the start, he concludes:
As the technology evolves and gen AI output becomes more accurate, we expect demand for our gen AI services and expertise to increase greatly over the next six to 12 months.
Over at Infosys, CEO Salil Parekh can point to having 57,000 employees trained up in generative AI skills, while the firm is currently working on over 90 generative AI programs. Arguing that the firm’s Topaz generative AI capablity set is helping boost market share, he’s keen to emphasize that the firm is working with both proprietary and open source Large Language Models (LLMs):
We are not developing a Large Language Model of our own. There are a large number of these models which are already in the market. Some of them are proprietary and some of them are open source. We are working with both types of those models.
He goes on:
Typically, we are working in what’s called the narrow transformer approach, which really we start to see data sets which are a little bit more enterprise-focused, which allow a large client to take advantage of that data set for their own activity. We are seeing applications on software development, on text, on voice, on video. We are seeing applications today on all of these areas, actually working on all of these areas and that is for the clients. Then, we are doing some work inside Infosys as well for our own activities.
Meanwhile at HCLTech, CEO C. Vijayakumar cites “hundreds of opportunities in emerging areas like gen AI”, some of which have already converted to wins and are in the implementation phase. The firm is also ‘eating its own dog food’ with internal generative AI adoption, he says:
We are leveraging internally gen AI across all our corporate functions, primarily to improve our employee experience and productivity. Various programs are in pilot phase across our people function, finance, risk and compliance, and sales and marketing functions.
And from a client-facing perspective:
We're working on generating early-stage opportunities as well as training our delivery organization to leverage gen AI for core development, deployment, testing, and managed services. Our software offerings continue to embed AI into their product offerings. I believe it's still early days as clients are still evaluating them as [proof-of-concepts], while ensuring their data strategy is right, as well as establishing the guardrails to ensure data governance and compliance requirements.
Vijayakumar offers up a number of use case examples:
We expanded our partnership with a Fortune 50 communication services provider to transform its billing operations, superior activation through billing experience to its customers. This would be done by leveraging advanced digital technologies, including gen AI.
One of the largest global technology companies has selected HCLTech as a preferred partner for their product developer support lifecycle optimization by using AI and Machine Learning. We will help the client deploy NextGen AI technology across various product cycles to result in faster triage and improved product quality and enhance developer ecosystem satisfaction.
A Europe-based financial services firm selected us to leverage gen AI to strengthen its global compliance framework. We will deploy large language models, along with machine learning technology for proactive surveillance and risk detection.
As with Accenture, there’s a lot of activity and clearly a perceived revenue stream being vigorously pursued here. Equally there’s the all-important and essentially unanswered question of when that revenue starts to make significant impact on the balance sheet.
TCS COO N G Subramaniam says that the revenue situation is “neutral” at present:
I think we are in an investment cycle like our clients. We as a company are also investing a lot in training our employees, in creating learning platforms and separate environment, like…the AI Playground which we launched, in which we are making available all the digital infrastructure tools and technology in which people can train and people can actually develop things and get used to how this particular technology works and that can be meaningfully deployed to our clients.
There's a lot of investment happening outside. Our clients also invest in the technology to do small projects, proof-of-concepts and proof-of-values, and then structure typical use case and so on and so on. We've also invested in cataloging several use cases by verticals and by technology to see how this particular technology can meaningfully deliver value on a sustainable basis.
But for now, the conclusion is:
I think in all that, there is demand, but I wouldn't rate it to be very high. But I would say that I think all directions indicate that this is something that will mature, and the potential for embedding it and embracing it all across our IT services value chain is huge.