How Shreeya Rashinkar Turns AI Skeptics into Agentblazers, One Telco At a Time

“Oh wow — so this really does work.”

It’s a reaction Shreeya Rashinkar hears often these days. As global head of presales at Tech Mahindra and a veteran Trailblazer, she’s been leading digital transformation projects for more than 18 years, partnering with Salesforce customers to turn legacy processes into new, smarter ways of working. Today, she leads the charge on all things Agentforce, helping global telcos launch production AI agents that handle everything from billing and HR to customer service and sales. 

We sat down with Shreeya to talk about her approach to change management, the top lessons she’s learned from launching Agentforce projects, and her passion for empowering the next generation of Trailblazers.

What made you choose the Salesforce ecosystem as a career path?

I think I was destined to do Salesforce. I started my career with Siebel CRM and then transitioned into transformation programs. While managing a project that was Oracle-heavy with a Salesforce CRM on top, we took over from a system implementer. That whole project came under me, and I was thrown right into the deep end. At that time, I didn’t know about things like Apex governor limits, but I had to get my hands dirty to fix a lot of issues and get the project to a successful state. It was a good experience that taught me a lot.

What does a typical day look like for you as a global head of presales at Tech Mahindra?

I head global presales and solutions for our telco customers at Tech Mahindra and I lead all our AI initiatives across various verticals on Salesforce platforms. On a typical day, my job includes talking to our telco customers and helping them get the most value from their Salesforce investments. There’s always new features and upgrades and a lot of customizations in these systems and since I understand how to get the best value out of it, I act as the customer’s ally and go-to person for guidance and best practices. 

I also manage RFIs and RFPs for digital transformation, managed services, and vendor consolidation projects. To do that effectively, it’s important for me to understand all the products, from Marketing Cloud to Communications Cloud, and how they can help our customers. 

What are some of the biggest challenges your customers face that you help solve with Salesforce?

I come with almost 18 years of professional experience, starting as a developer on Siebel CRM before moving to Salesforce. On the telco side of the world, the products themselves are really complex. We’re talking about internet connectivity, broadband services, OTT services, and mobile services. The challenge is ensuring that Salesforce can act as a powerful orchestration engine to manage things like order fulfillment, dynamic pricing, and contract rollouts.

The telecommunications industry is complex because of the sheer volume of orders we have across B2B and B2C. It’s a very interesting area to be in because of the complexity of systems and integrations on the BSS and OSS stacks. 

How has the emergence of agentic AI changed the way you approach these challenges?

This is a very important question. I recently had the opportunity to showcase our Agentforce go-to-market solution, which we specifically designed for telcos, to our Tech Mahindra board. I was very excited to show them how AI could fit in and how the story actually differs for each customer.

It’s important to bring customers along on this journey from predictive to generative AI, from chatbots to AI agents. A chatbot is something with a predefined set of instructions — you can only do as much as you’ve configured. Agentic AI, on the other hand, is designed to be available 24/7. It doesn’t have to be pre-configured for every corner scenario — it can automatically take actions and decide what to do next.

During one of our discussions someone threw out a hypothetical scenario for our AI Agent: “I disconnected my mobile services like two months ago, but why am I still getting billed for it?” The agentic AI understood that the customer sentiment was negative. It went ahead and automatically created a case in the system with the right type, subtype, and priority. We even had a Slack agent that would create a special channel, saying, “This is an important customer, they’re unhappy, take care of it immediately.” This worked flawlessly and was very convincing.

How have customers been reacting to Agentforce?

Most customers are initially skeptical, so it takes a lot of groundwork to convince them to go to the next level. During our initial demos, we’ve heard reactions like, “Oh wow, so this really does work.” That was the starting point where we knew customers were getting interested. A challenge we’re seeing currently is that many of the large telco enterprises have their own LLMs, so that’s really where it becomes even more important to talk about all the native capabilities within Salesforce and the advantages of building with Data Cloud. We’re seeing a lot of excitement with Service agents. In production, we’ve seen at least 30 to 40% call deflection so far. 

As a Salesforce veteran, how do you stay current and upskill yourself with these new technologies?

I always try to take a hands-on approach. With generative AI, for the last two years, I made sure I wasn’t just reading about it, but I was actually doing it. That makes all the difference. Because I come from a presales and sales background, I have to convince myself first before I can convince a customer. That’s why I’m always upskilling myself, using resources like the partner pocket guide and the Simple Demo Orgs (SDOs), where you can try whatever you want.

Initially, this was a personal ambition, but then I realized the power of this technology. I did a lot of enablement sessions internally within Tech Mahindra, where we have more than 2,500 associates dedicated to Salesforce. I thought it was important to percolate this knowledge to a wider audience. We also held a “Precursor to AI Now” event across our Tech Mahindra campuses of Mumbai, Pune, Delhi and Bengaluru along with the local trailblazer communities and a special “AI Now” tour event with Salesforce at our Hyderabad campus. These events are designed to build more awareness and fluency around things like Data Cloud, prompt engineering, and agents. As a result of these initiatives, we have more than 1,000 people who are already AI Associate (1100+) and AI Specialist (400+) certified.

What’s your biggest piece of advice for someone just starting out with Agentforce?

My advice is to always keep your audience’s unique context in mind. If a person is working as a developer, you should talk to them about the tools that speak to their level of technical depth, like Flow. That is where they’ll see the value of it. If you can cater to your audiences with the right use cases, you can get them to be a part of the whole AI ecosystem. Now that the Developer Editions are available, there is absolutely no reason why somebody should not be learning about Agentforce.

What has surprised you the most about working with Agentforce?

I’m most surprised by how quickly the product is evolving. When I’m using the SDO orgs, I get to see the daily updates before they are publicly released. For example, I was able to try out the sales coach agent before it officially came out. The product teams have an amazing vision and a detailed plan for what’s next. 

I’m looking forward to trying new features like the voice part, which I haven’t done yet, and the multimodal functionality, which we have already put a use case around. I’m also excited to explore MCP and how that will redefine the way we can talk to external systems like payment gateways.

How do you approach change management and convince customers that AI won’t replace their workforce?

This is a very common concern. I was once presenting to a B2C customer who asked me, “We understand the good part of AI, but if our teams become more efficient, how do I utilize them now? I can’t let them go.”

My response was that AI doesn’t replace people — it empowers them. With the time they’ve gained back, they can focus on more strategic work. They can tackle the massive technical debt from years of customizations, create new business roadmaps, and proactively upgrade the system with new features from Salesforce’s releases. AI eases their workload, allowing them to do a better job at ensuring the system is more scalable and stable. This approach convinced that customer and they’re now including Agentforce in their new service rollout.

What do you enjoy doing in your free time?

I try to relax. Because of all the continuous meetings and things happening in our daily lives, we don’t always pause. So I try to relax a bit, and I enjoy a little bit of reading and meditation.

Shreeya’s top 3 Agentforce learnings 

1. Testing is paramount. Since agentic AI is not predictable, you need to do a lot of testing. We currently use spot checks and have a few people putting agents through their paces to find corner scenarios. This is a key learning we’ve taken to make sure the agent is robust.
2. Define clear instructions. We’ve seen that if there are multiple people using the agent in parallel, it sometimes doesn’t give the right answers or does things it’s not supposed to do. We are still getting the hang of this by making sure the topics and instructions we are defining are very clear.
3. Choose the right ROI-driven use cases. We have to decide whether a normal automation would work versus whether we really need agentic AI to do a task. Since Agentforce comes with its own cost, we have to make sure the use cases we automate are truly creating a significant ROI for the customer. People will question, “Why do we need something so costly when we can do the same job with some Python code?”