Akshat Saxena describes himself as a student of philosophy as much as technology. It is an unusual thing for a startup founder to say, and it shows in how he talks about his company.
When asked about artificial intelligence replacing workers — a question he must hear constantly — Saxena pauses and says he does not know how to answer it as a technology person. Then he tries anyway.
“AI, by definition, is an intelligence layer,” he told Business Outreach. “Think of AI as an enabler for a human workforce. Maybe the same sales executive who goes out tomorrow to the same retailer now has greater capability to position the best product.”
Saxena is the co-founder and CEO of Vibrium AI, a Gurugram-based enterprise agentic AI platform he founded in July 2025. The company builds what he calls an agentic operating system for large organisations: autonomous AI agents that can reason, take multi-step actions, and optimise processes across industries including BFSI, e-commerce, retail, and SaaS.
AI in enterprises and where Vibrium comes in
The problem Saxena is working on is the absence of AI that works at scale in enterprises. Most large companies in India have a pilot running somewhere. Most of those pilots go nowhere. IBM research found that while 59% of enterprise-scale organisations in India had AI in active use, 27% were still only exploring it, and the top barriers to scaling were skills gaps, the absence of suitable platforms, and difficulty integrating AI across existing systems.
Globally, McKinsey’s 2025 workplace report found that only 1% of leaders describe their companies as mature in AI deployment. The global enterprise AI market stood at $30.18 billion in 2025 and is projected to reach $570.36 billion by 2035, growing at a CAGR of 34.2%. The opportunity is enormous. The execution rate? Not so much.
Vibrium’s platform sits in the space between the large language models built by the likes of Anthropic, Google and OpenAI, and the business outcomes enterprises actually need. Saxena is deliberate about this positioning.
“Likes of Google and Anthropic and OpenAI, they are investing into what we call foundational models. Think of them as intelligence highways. They are not the end solution in themselves. For you to bridge that gap from a foundation model to an end solution, there are a bunch of layers that have to be added in between.”
The tech entrepreneur breaks it down using a food analogy to explain what those layers look like in practice. “Think of Vibrium as an AI cloud kitchen. If you have a recipe, you can create a biryani or a pizza or a burger. The cloud kitchen gives you all the ingredients, all the tools. You create your own dish. And that dish, in the case of an enterprise, is the solution to a specific business objective.”
Inside a beam-and-pillar model
The platform itself is structured on what Saxena calls a beam-and-pillar model. The horizontal beam is the core infrastructure: LLM integration, process intelligence, automation layers, compliance guardrails, real-time monitoring, audit trails, and adaptive learning. The vertical pillars are the industry-specific deployments built on top of it.
An HR department can deploy agents covering the full employee lifecycle from recruitment to exit. An FMCG company dealing with daily sales force absenteeism can replace lost field coverage with AI agents that reach out to retailers, pre-build orders, apply promotional schemes, handle objections, and cross-sell.
What happens when an AI agent gets it wrong? Saxena frames the answer around knowledge provenance. Vibrium’s agents operate on the enterprise’s own knowledge base: its products, pricing, processes, and systems. If that knowledge base has gaps, the agent’s output reflects those gaps.
“The constraint is not technology. The constraint is knowledge, and knowledge is typically supposed to be provided by the client,” says Saxena.
Revenue-funded and expansion-ready
Vibrium closed a pre-seed round of close to $1 million at launch and has since been funded primarily by client revenue. Saxena describes the company as financially disciplined and says revenue is scaling at the same rate as the company itself.
Talking about expansion, the geographic plan runs in two phases. Middle East and Southeast Asia — UAE, Saudi Arabia, Singapore, Indonesia, and Vietnam — within three to six months. The US market follows in Q4 this year or Q1 next year.
Saxena’s five-year target is to help client enterprises deliver three times better outcomes at their current cost base, and to contribute $15 billion in measurable client value over the next three to four years.
India’s enterprise AI adoption rate, at 83% of surveyed enterprises in IBM’s research, ranks second globally, behind only China. The demand environment Vibrium is entering is already in place.
“We have a potential partner in the form of AI that can allow the world to continue to run without necessarily having to invest as many man-hours into that. And thereby those man-hours that are released — you give them back to humanity. Maybe give them an option to spend that time with their family, or doing what they love,” Saxena concludes.
