At India Mobile Congress 2026, Yogendra Shukla, Co-founder of NetPrompt, spoke to Tech Achieve Media about the evolution of agentic AI, the challenges of moving from pilots to production, and the need for secure, deterministic AI frameworks for enterprise IT infrastructure. He also discussed NetPrompt’s approach to infrastructure automation, its partner ecosystem, and why India needs to strengthen its product-building mindset.
TAM: NetPrompt is a relatively young company. Could you tell us about its journey, the problem it is trying to solve, and its offerings?
Yogendra Shukla: NetPrompt is a two-year-old company founded by Pradeep Rasamsetty and me. We met at Y Combinator, where our journey began. Pradeep, who previously worked at Cisco, was looking for someone to lead go-to-market (GTM), sales, marketing and the business side. I come from a technology sales background, having previously worked at Nokia. Our complementary skill sets worked well together.
We started working together at the pre-MVP stage. While Pradeep focused on technology, I worked on positioning the offering and identifying the right approach to taking it to market. This was different from traditional B2B solution selling, where one typically works with established accounts. We experimented with different approaches, including channel partnerships, and began building connections with companies such as Redington, IBM and Tata Consultancy Services.
Over the past year, we have gradually developed this ecosystem. We participated in the India AI Summit, and Gartner mentioned NetPrompt in one of its reports in January. We were also featured in a July report on innovative vendors in agentic AI. These mentions provided validation that we were working on a problem the market needed to address. We have also engaged with NASSCOM, presented at its GCC Summit and received an opportunity to showcase our innovation at India Mobile Congress 2026. NetPrompt is an India-founded company based in Bengaluru. Pradeep leads technology, while I focus on the business side.
Our focus is on deterministic AI for IT infrastructure, including data centres, network operations centres (NOCs), manufacturing enterprises and other environments where security, reliability and governance are critical. In such settings, organisations cannot simply connect their live environments to open models and allow them to operate without adequate safeguards.
We are not trying to build yet another AI model. Our objective is to use existing models and make them commercially viable for industrial adoption. NetPrompt provides a deterministic AI and governance framework that incorporates security and supports multi-vendor devices. It does not replace existing tools; it enhances their capabilities. Globally, companies such as Selector, Itential and NetBrain operate in related areas. NetPrompt represents India’s attempt to build a comparable framework for secure, governed infrastructure automation. We are available on AWS Marketplace and Google Cloud Marketplace, and we are looking to expand our alliances further.
TAM: Agentic AI was initially seen by many enterprises as an emerging technology to experiment with through pilot projects. Has the scenario changed, and how do you see the technology evolving?
Yogendra Shukla: Yes, the scenario has changed. The current wave of AI adoption gained momentum with the commercialisation of generative AI, particularly through products such as ChatGPT. Initially, the interaction was largely question-and-answer based: users asked a question, and the system generated a response. That led to a broader question: should AI remain a question-and-answer machine, or can it be assigned a specific task and allowed to carry it through to completion? Agentic AI is an attempt to move beyond the chatbot model towards systems that can act within a defined environment to accomplish a task.
For example, when an IT service management (ITSM) ticket or customer support ticket comes into NetPrompt, we do not simply receive it and provide an answer. The system can perform root-cause analysis, identify a remediation path and seek human approval. Once the approval is granted, it can proceed with the fix and report the action taken.
That is the distinction between a chatbot and agentic AI. The latter moves towards taking action on a problem rather than stopping at a response. I believe we are still at the beginning of this journey. Artificial General Intelligence (AGI), which aims to bring AI closer to human-like reasoning across a broader range of tasks, represents a further stage in this evolution. There are already efforts to advance towards that goal, although its timeline remains uncertain. My view is that 2026 marks an early stage in the adoption of agentic AI. We should see wider adoption through 2027 and early 2028, with agentic AI potentially becoming a standard component of enterprise automation. We may also see initial pilots exploring more advanced forms of general intelligence.
TAM: Industry discussions frequently point to high abandonment rates for agentic AI pilots. Why are enterprises struggling to move these projects into production, and what can startups do to improve conversion?
Yogendra Shukla: One of the key reasons is the nature of the B2B procurement cycle, which is often long. Startups need sufficient runway to survive that cycle and continue engaging with potential customers. A pilot is relatively easy to initiate because it allows an enterprise to explore a technology with limited commitment. The situation changes when the discussion moves to a paid subscription or a larger commercial engagement. At that stage, management begins asking whether the technology is an immediate business requirement or something that can be deferred to the next financial quarter. That is where many pilots lose momentum. The technology may be interesting, but unless the business problem is urgent enough, the organisation may not prioritise the investment.
My advice to startups is to adopt a signal-based approach to marketing and customer outreach. Instead of approaching every potential customer with the same pitch, identify organisations that are already experiencing the problem your solution addresses. Research the signals that indicate a need, understand their pain points and approach them with a relevant proposition. This can improve the chances of converting a pilot into a commercial engagement.
Networking is equally important for startups. We are still part of that journey ourselves. One useful approach could be to build stronger startup networks and consortiums, where founders at different stages can share experiences and help one another navigate challenges. For example, NASSCOM has cohorts that connect startups across different stages of their journeys. An earlier-stage company can learn from a more mature startup that has already addressed similar challenges. Rather than reinventing everything, startups should learn from existing experiences, build relationships and use those connections to move forward.
TAM: NetPrompt connects AI with existing IT infrastructure. What foundational capabilities should enterprises have in place before introducing agentic AI?
Yogendra Shukla: Enterprises need well-defined policies and a reliable operational structure. They should first ensure that their existing infrastructure and processes are functioning properly without automation. Automation should not be introduced as a quick fix for an environment that is already broken. If the underlying processes are unreliable, automation will build on those weaknesses and can only deliver limited improvements. Organisations should establish clear operating procedures, governance policies and reliable workflows before introducing agentic AI. Once the foundation is stable, they can identify suitable processes for automation and gradually expand their use of AI.
TAM: Several companies are now offering agentic AI solutions. What differentiates NetPrompt from other players in the market?
Yogendra Shukla: The answer depends partly on the sector. Many agentic AI companies are focusing on financial services and banking, financial services and insurance (BFSI). We are concentrating on IT infrastructure, which has its own requirements around security, governance and deterministic outcomes. In our conversations with large enterprises, we have found an interest in collaborating with us. These discussions have encouraged us to believe that we are addressing a relevant requirement in the market. We are still a young company, and we see collaboration as an important part of building the business.
There are several established international players in related areas, and competition is healthy because it pushes companies to improve their offerings. Selector provides an interesting example. Rather than treating Itential purely as a competitor, Selector partnered with the company, and the two have subsequently worked together on a major deal in the US. That demonstrates how companies can collaborate to address customer requirements instead of viewing every market participant only as a rival. We believe there is value in taking a similar approach. By building partnerships and working with the broader ecosystem, we can strengthen our offering and expand our reach.
TAM: Finally, what changes are needed in India’s technology ecosystem for the country to become a stronger product innovator, particularly in emerging areas such as agentic AI?
Yogendra Shukla: I believe two things need to change in India. First, our students and aspiring entrepreneurs need to become more comfortable with failure. From my experience in international sales and business development, I have seen how important it is to overcome the fear of rejection. People should be able to experiment, build new ideas and pursue opportunities without feeling that failure carries a stigma. That mindset is essential for innovation.
Second, we need a stronger product mindset. As Jyotiraditya Scindia also highlighted during the event, India needs to think of itself as a product-building nation rather than primarily as a services-led economy. For historical reasons, we have developed considerable strength in services. However, we must also invest in building products that can compete globally and establish India as a leader in technology innovation. When we start building products at the forefront of technology, we move towards becoming a technology-leading nation. Agentic AI is one of the areas where Indian companies have an opportunity to contribute to that future.
