InferKNOW co-founders Chandraleeka Hazarikka and Shivang Ajay Desai explain the contextual gap they see in modern web architecture, how they sharpened the category story for the 2026 Pitch Competition, and why agentic commerce is next.
About Chandraleeka Hazarikka and Shivang Ajay Desai
Chandraleeka Hazarikka and Shivang Ajay Desai are co-founders of InferKNOW. Their team brings four startup exits, experience co-building a unicorn, and more than 40 years of deep-tech experience. They are building a real-time context layer intended to help web systems interpret session intent before pricing, fraud, personalization, and agent workflows act.
What problem are you solving, and what convinced you it was the right opportunity?
InferKNOW is solving what we call the contextual void in modern web architecture. Nothing in the modern stack reads the live moment.
Companies have invested heavily in knowing who is on their site: CDPs, analytics, identity and risk engines. But none of them can tell you why that visitor is there right now, or what the session is worth. That gap is expensive. Teams default to blanket discounts, generic journeys and false declines because their systems act on stored profiles, not on the moment.
InferKNOW is the context layer in between. It reads each session at the edge, in milliseconds, and turns existing signals into structured context and intent for pricing, fraud, personalization and agents to act on. No stored profiles, no identity required, no re-platforming.
What convinced us was the evidence. In our company-reported results against a holdout group, we've seen +200% purchase completion and +57% yield per session. We also saw how fast it goes live: a single script, about four hours of engineering, live in 48 hours. When a layer is that easy to adopt and measurably moves revenue, the opportunity is clear.
In one of our company-reported early enterprise deployments, a customer had to ask us to shut down the software for a few days because the spike in sales was so sharp that they were running short on inventory because the demand had not been modeled.
How have your previous ventures shaped how you spot opportunities, and what's different about building this one together?
Our team brings four startup exits, one unicorn co-built, and over 40 years of deep-tech experience. That experience taught us to look for infrastructure gaps: places where every company is building the same workaround because a layer is missing. We also learned that the best infrastructure fits into what customers already run, which is why InferKNOW runs alongside the existing stack.
What's different this time is that we've worked together before, so we skip the trust-building phase and go straight to the hard problems.
As founders, one of us leans technical and the other commercial and brand, which means product decisions and go-to-market decisions get made in the same conversation.
What are companies overlooking as AI assistants and agents shape discovery, and what should they prepare for?
Companies are still optimizing for human visitors with identities and profiles. Increasingly, the visitor is an AI-referred user arriving from an assistant's recommendation, or an autonomous agent acting on someone's behalf. Those visitors often have no cookie history, no login and no stored profile, and a lot of existing tooling can't tell them apart from humans.
The traffic is already here. Adobe reported in August 2026 that AI-referred visits to US retail sites rose 62% year on year in July and converted 60% better than other traffic. The July 2026 launch of Digital Commerce 360 and ReFiBuy's AI Commerce Rankings reported that 87.1% of the top 1,000 US retailers already received measurable AI traffic, with average readiness of 42 out of 100. Sources are linked below.
Companies should prepare to tell humans, AI-referred visitors and agents apart in real time, because each needs a different response. They need to understand situation, not just identity, because an agent's identity says little about what it's trying to do right now. They also need machine-readable context and terms, including structured offers and agent-to-agent negotiation, so their systems can transact with AI shoppers and booking agents.
Autonomous systems can act only as fast as they can understand. The winners will supply the context before the system acts.
What was the most memorable part of the AI Builders Global Conference 2026, and what did preparing teach you about communicating your vision?
The talks and opportunities to network with other AI builders and ecosystem experts were exceptionally memorable. Preparing forced us to cut the story down to one idea: your systems know who is here, and InferKNOW knows why, right now.
Infrastructure is easy to over-explain, so we learned to lead with the problem and the measured outcome, and to show the Who -> Context -> Intent -> Action flow in a single picture. We also learned to say what we are not: not a CDP, not personalization, not CRO or analytics, but the context layer that feeds them. That makes the category clear faster than a long feature list.
What's the next major milestone, and what opportunities excite you most?
Our focus is bringing the 48-hour deployment to more enterprises and verticals where the context gap is costly: retail and eCommerce, financial services, fraud and risk decisioning, travel and hospitality, and media.
The opportunity we're most excited about is agentic commerce. As AI shopping and booking agents become real buyers, there will be a need for live, machine-readable context and for structured terms between agents. We see InferKNOW as the layer those systems rely on before they act.
What would you say to founders considering pitch competitions, and why do communities like AI Builders Network matter?
Do it. A pitch forces you to find the one sentence that makes your product clear, and it gives you feedback from people with no stake in being polite. Prepare for the questions, not just the talk, and lead with the problem and proof rather than the technology.
Communities like the AI Builders Network matter because building is lonely and fast-moving. They put you in front of other builders, potential customers, investors and mentors, and you hear what the market actually cares about.
For us, the AI Builders Conference led to a conversation about deploying InferKNOW as the context layer for feeding existing SaaS applications, and it is something that is now actively moving towards a commercial deployment.
Sources and context