AWS bets on GenAI, agentic tech and skilling to power India’s cloud

On the sidelines of the AWS Summit event 2025, Bengaluru, Shipra Sinha, Lead Analyst- Industry

Intelligence Group (IIG), CyberMedia Research (CMR) sat down with Satinder Pal Singh, Head of Solution Architecture, AWS India and South Asia

Satinder Pal Singh, leads a team of architects dedicated to assisting customers in designing scalable and efficient solutions using AWS services

How would you define the evolving role of a Head of Solution Architecture within the fast-changing cloud landscape of India and South Asia?

We are committed to provide the power of AWS directly to our customers and partners so they can be successful in what they are doing. They should be able to get all the benefits of AWS services, develop their workloads and their architectures so that they're successful. Secondly, we are going through a very transformative time and Gen AI is an era of discontinuous flux. The kind of potential that Generative AI brings can have a transformative impact to our customers. So, we are invested in ensuring that customers are aware of the transformative impact. What can we do to demystify and help them build Generative AI applications in a secure and responsible fashion that is one of the big parts of the role. The third one is always thinking about inventing on their behalf. What are the things that they are telling us and work backwards from there, but we also decipher what are the things that they're not telling us. So, thinking about as we call as a divine discontent that customers would have, that even when they say that they are satisfied, there is an element of dissatisfaction. Uncovering that and finding out, how can we innovate on their behalf is the third big thing about the role.

What are your key strategic priorities for accelerating cloud adoption and driving innovation across diverse industry verticals in this region?

Broadly, they are classified into three buckets. The first one is developing industry solutions and putting in things that have worked for a specific industry and the technology architecture behind it. An example is Pocket FM, an AWS customer who is building audio series platform. The content production team has created a story and wanted to scale rapidly. They wanted to leverage AI to see whether they can convert that story, make it suitable for other regions as well, other areas, other countries. The team understood their requirements and used AWS Generative AI technologies so that Pocket FM can rapidly expand to 35 plus countries. That's the first thing that we're doing.

Secondly, we work on ensuring that our customers responsibly build AI application. They know what are the architecture patterns. We advise them on what changes are going to happen and give them an early visibility into what are some of the trends. One big example of that is Agentic AI. If you look at Amazon Bedrock, which is our Generative AI solution, we introduced agentic capabilities on Amazon Bedrock, which is multi agent collaboration. Multiple agents can coordinate to deliver a task which is a very powerful capability. So, ensuring that we give customers a framework and advice so they can harness the capabilities of these technologies.

The third one is skilling. We have trained 5.9 million people in India on cloud technologies specifically for AI. We've got AI Ready program and we planned to have over 2 million people trained on this by 2025. This commitment became a reality faster because we reached two million people at the end of 2024, one year ahead of schedule. It's a free program and consists of eight modules. The modules are diverse. So, skilling is very important so that the when we got the right technology, we need to ensure that we have trained people who can harness the capabilities of these technologies.

How do you foster architectural innovation across both established enterprises and digital-native startups with differing needs and maturity levels?

The foundation of this is working backwards from customer. With 240 services, which consists of services which are infrastructure to manage services, etc, we've got set of services that suit requirements of different kind of customers. So, working backwards from customer requirements is the most important thing so that we understand what their needs are and how we can architect. Let’s talk about Zeta. They were trying to disrupt core banking; they reimagined what their application would look like. So, we helped them provide the right kind of architectural constructs and tell them that this is the way you can harness abilities of cloud. These are how you'll take care of security regulations, security controls so that they can achieve what they achieved.

Second example is about an enterprise customer, Apollo tyres. The challenge that they had was their equipment spread out in various countries. They were unable to get a centralised view on equipment effectiveness, how they can better utilise that. We worked with Deloitte, which is one of our partners to create a solution which is based on our IoT technology, which gave them a centralised view that improved the productivity by 9% at that scale resulting in massive cost saving. That's another example of coming out with a partner, building an industry solution and helping them solve a business challenge. So, regardless whether a customer is looking at a solution or customer is looking at a specific need of a service, addressing scale, performance, etc, we provide them the right kind of opportunity.

What role does the broader partner ecosystem (including startups, system integrators, and ISVs ) play in shaping AWS’s architectural success in South Asia?

Partners are a very important capability for AWS. We work with partners to create the right kind of solutions for customers. We have assigned competencies to partners so that we know that they are capable and they have proven credentials. As an example, we have Shellkode, which is a Gen AI competency partner. Shellkode has developed multiple solutions for customers. This gives customer confidence that if you are engaging with the partner, they know the partner is capable. Technology competency, industry competency, we accredit partners across these competencies and then work with them. So, partner is a very important and integral element, because they'll bring the whole ecosystem and help customers gain the benefits.

How do you maintain architectural consistency and scalability while supporting a diverse customer base and complex partner landscape?

We have a framework which is called AWS Well-Architected Framework. That framework was developed based on tremendous amount of learnings that we got working with customers. The framework brings out what is the best way that you can architect applications and workloads on cloud and then it breaks it down into six categories that what would a good architecture look like.

The first pillar is Performance Efficiency, second pillar is Reliability, the third pillar is Security, fourth pillar is Operational Excellence, fifth is Cost Optimisation and the last one is Sustainability. This is the framework that is available to all customers and to our team so they can evaluate their architecture based on these six pillars and see whether they are adhering to some of the best practices or not. So, it's a very powerful capability and we have exposed that in an automated fashion, it’s called well architected review tool. It's available for customers also so that they can self-evaluate as the tool is available at zero cost to them.

What are some of the unique architectural or regulatory challenges you encounter in South Asia, and how is AWS addressing them?

The kind of infrastructure that we have put in India, we have Mumbai region, we have Hyderabad region. There are three availability zones in each region, a total of six availability zones. If you look at the power of that, we can build and deploy any applications, regardless of what the needs would be. An example of that could be, disaster recovery. There are some customers who would need their disaster recovery site to be geographically distributed. Some of them will say 300 or 400 kilometres or different size etc that’s what multi region architectures allows to do so.

The second element that is very important, as far as regulating both the regions, is the data remains in India. Data does not leave India unless customer opt out of this. That addresses some of the regulatory requirements. The third one is security. We have hundred plus security services that customers can use. There are best practices, there are artifacts that are available. Security again consists of security of the cloud and security of workloads in the cloud. Security of the cloud is AWS responsibility on behalf of our customers. For security of their data and workloads in the cloud, we provide them 100 plus capabilities. Then there is a market list, they can look at security partners so that they can architect the solution in the right fashion that meets the regulated requirements.

In your view, what technologies will define the next wave of cloud architecture in India and South Asia over the next 3–5 years?

The one that needs a special mention of is Generative AI. It's going to have a huge and transformative impact on India and we are absolutely focused on driving that for India. An example is of IndiaMART. It is one of the world's largest B to B market place and they had this unique challenge that almost 60% of their customers used to come from tier two, tier three cities where english is not their native language. Using our Generative AI technology, which is Amazon Bedrock, they launched a property called hindi.indiamart.com which has 5 million SKUs and product listing available and they were translated giving the cultural context. That's one example of what Generative AI does.

The second example is Zomato. Zomato's restaurant partners do not have access to sophisticated equipment that they can click a really nice picture and send it over to Zomato. Zomato has a public policy that they will not use AI generated image because that's a false image. They use our AI technology and our purpose-built chipset, Inferentia to actually make that image look more appealing. So, they empower their restaurant partners who do not have access at scale.

The third example is DTDC. They launched an engagement platform called DIVA 2.0 that completely changed the way they interact with their customers, providing them information, using agents to take actions.

Generative AI is a sub part of artificial intelligence and Machine learning is an overarching term. An example is a very large fintech customer, does their KYC on AWS. KYC means that some sort of an identification document is produced and we need an artificial intelligence technology, not Generative AI technology, to take out relevant information. Then AI technology should be able to detect a real and valid identity.

Machine learning plays a very important role because there are customers who would want to develop their custom models. An example of that is Gupshup. Gupshup is a conversational AI platform. They used our machine learning technology, which is Amazon SageMaker, to train their models, build their models to solve the business problems that they were dealing with at 30% lesser cost as compared to building it. So, Amazon SageMaker happens to be a machine learning platform that helps customers leverage machine learning capabilities in an easy-to-use fashion. Amazon SageMaker Studio is making it easy for machine learning practitioners to use that technology and remove undifferentiated heavy lifting even for machine learning.

How do you foresee AWS role in driving digital transformation especially in India and South Asia?

There are two important elements. The first one is the fact that we need to continuously innovate on behalf of our customers, providing them the right kind of tools, technologies that they can solve the business problems. It's not something that we'll do today and pause it, this is a continuous exercise. We need to understand what are the challenges that they are facing. The second one is skilling which is a very important element. We need to skill the workforce to have capability that they should be able to harness. AI Ready Program is one big step towards it.

Which architectural trends do you see as most impactful in shaping the future of cloud infrastructure in India and South Asia?

A very promising thing is agents. Agentic AI has capability of simplifying and there are two perspectives. One is that it simplifies things for Generative AI developers for developing Generative AI solutions. Amazon Q for Java, for VMware, for .NET are agentic implementation that are helping customers develop or modernise their applications very fast.

The second one is agents the customer will use to solve a business problem for themselves. Multiple agents coordinating with each other and that's why multi agent capabilities of Amazon Bedrock will play a very vital role. So, agentic implementation will have a very big and transformational impact to business.