I spoke with Bharath Vasudevan, Chief Capability Officer, Tech Mahindra Business Process Services, on the company’s investments in GenAI, the business challenges it addresses, and the longer-term effects on CX services.

What are some of the immediate opportunities for using GenAI?

Bharath Vasudevan (BV): Like every other technology disruption, at Tech Mahindra we look at the business outcomes we are trying to solve and whether we are solving them for customer metrics, employee metrics, financial metrics, or a combination of the three. Conceptually, we want to be very clear on the business outcomes and these are the first levers in any GenAI initiative.

The second thing we look at is the GenAI strategy as part of the larger ecosystem. Some of the large enterprise hyperscalers have both the intellectual and financial bandwidth to model LLMs by themselves. Then, there are the platforms and products that are consuming these LLMs and creating products and services; the majority of the technology players today sit in this space. And then there are organizations leveraging GenAI to impact one or more of their business outcomes; these are the early adopters taking GenAI services and products and figuring out how they apply them. And others are waiting for evidence-based applications.

Our play is clearly in being part of the ecosystem that helps create these service offerings out of the LLMs and take them into the market. We have looked at six areas:

• Customer and agent conversation
• Content creation and curation
• Integration with enterprise data
• Assurance, governance, and risk compliance
• Enterprise business processes
• Process excellence.

For each of the six areas, every time we deploy use cases, we want to be very clear on what is the business outcome. For example, we are looking at content because we believe there is a huge play of LLM in how content gets created and curated across industries.

When it comes to process excellence, we are looking at not just our clients but our processes for employee utilization and efficiency maximization to improve how we run our BPO work today. The opportunity is to leverage AI to improve productive time and make our employees much more efficient.

How do you prioritize with the clients across these areas?

BV: We have formed within the organization what is called a generative AI council, a cross-functional team of about 15-20 leaders from across the organization. These people are, in turn, distributed into pods with each pod focusing on a particular business area and identifying GenAI applications.

Which use cases would you highlight having the fastest route to implementation with a strong ROI?

BV: Some of the initial use cases that we are starting to develop are in customer experience. For example, for a large U.S. telecom. It has a tedious process to update multiple platforms, knowledge bases, automated workflows, and deployed chatbots. These frequent process changes are time-consuming and dependent on IT with the knowledge coming from disparate systems.

We are creating a common source of information and using GenAI to extract key sources of information and use it in a common database, not just for our agent conversations. We are also looking at conversations from the website and other channels not owned or influenced by Tech Mahindra today.

This GenAI solution will track and synthesize information to automatically update in dynamic nature all these platforms, digital properties, and internal systems to improve volume management, efficiency, and customer experience.

I would expect it will impact employee experience as well?

BV: Significantly. I just looked at the end outcome, but you know, eventually it has a direct impact on employee experience as well.

The main benefit is taking information available in any format from any channel and bringing it into a template or a framework that is common and every agent can use. Being able to take information from disparate sources in disparate formats is really where GenAI plays a significant role.

Is there an element of human supervision on top of this?

BV: We must look at machine learning and GenAI as an active process where you are constantly feeding the technology with new data, new data sources, new data types, and new data exceptions. Without having the human in the loop to manage those interactions and provide the technology with all the training that it requires, I do not think you are influencing the ability of the technology to learn as significantly as you should. None of these technologies is taking away the human in the loop.

A classic example is with one of our largest telco accounts, a U.K. telecom. We have a significant amount of chat as a channel in the U.K., and the question is to standardize and personalize these responses using GenAI to serve as agent co-pilots during live interactions.

Given the maturity of this account, we have moved significantly beyond FTE-based pricing. This is an account where almost everything we do for the client today is tracked and measured by outcome. The use cases we are building for this client are also structured more towards customer experience rather than just efficiency.

Today, we will be using GenAI for support scenarios, but eventually, we will also use this to impact upsell and cross-sell.

What stage are you at with deployment?

BV: We are at a stage where we are creating smaller POCs from the floor and then starting to present them to the client. I think we should be in a position to start showcasing some of these tools by the end of 2023.

You mentioned GenAI use cases on the marketing and content side. Can you expand?

BV: One is to leverage unstructured data from external marketing sources to provide insights for sales enablement for better GTM and business development initiatives.

Another is using GenAI to generate personalized content based on mining CRM data to enable the sales function to improve customer insights and accelerate sales. It is largely focused on B2B and we will pilot this internally at Tech Mahindra, given the amount of sales and marketing activities the organization engages in.

One of the goals is to make the RFP process a lot more effective by extracting maximum knowledge from whatever we have in our database. We are also working on creating an RFX factory using GenAI to complete 80% of the straight-out-of-the-box information for the solution architects to solve the remaining 20%. Yet another is input gathering from large contract reading with GenAI.

Most of these are speed and quality enablers, which eventually also improve our outcome metric in terms of winnability and our ability to execute on client engagements.

These marketing and sales use cases are further drawn out into early 2024.

How are these new technology capabilities impacting clients' outsourcing decisions?

BV: When they look at the future and at strategic partnerships, they want to be working with vendors that can provide them with a richness of perspective and a vision for the future, including in their own industry from a CX viewpoint.

Whatever GenAI solution we propose should help create a multiplier effect to what they are doing themselves rather than compete with it. We partner with some of the largest LLM developers as a service provider and want to help them take these GenAI products and services out into the market and help evangelize and deploy.

What is the most exciting area emerging from the rise of GenAI in CX?

BV: The most fascinating part about GenAI is the human in the loop in the CX context. I am looking at the evolution of roles, responsibilities, and skills. I think that industries and companies that manage this evolution well and can clearly define the roles and responsibilities of the human in the loop, and can make that work for the business, are going to outshine the others.

It applies to executives like me. It applies to people managing projects. It applies to agents. It applies to everyone.