Company · About
Building the way brands sell in the AI era
Buying is becoming asking. Shoppers ask questions on your site, ask ChatGPT what to buy, and send AI agents to compare products on their behalf. Stores were built for clicking. Kinect was built for this.
We give every brand one intelligence — trained on its catalog, policies, and customers — that sells to the shoppers on its site today and represents it to the AI agents doing tomorrow's buying.
Why Kinect exists
Kratik and Varun met at Reevo, building AI that understood B2B sales conversations — why deals closed, which objections mattered, what signals predicted intent. Kratik led Conversational Intelligence; Varun built the context graphs underneath it.
The same gap kept showing up in ecommerce, and nobody was fixing it. Brands have a dashboard for every click and conversion, and almost nothing on what shoppers actually want. The search bar captures keywords and throws the intent away. The shopper with one unanswered question doesn't file a ticket — they bounce.
Kinect closes that gap. An AI sales rep grounded in the brand's catalog answers that question in the moment and sells the way the brand's best employee would. Every conversation becomes first-party intent data the brand has never had. And the same catalog is made readable to the AI agents shopping from outside.
What we believe
Intent
Intent over keywords
The search bar captures words and throws away what the shopper actually wants. Reading the intent behind every visit — and answering it — is the whole job.
Voice
Your voice or nothing
A generic assistant bolted onto the page erodes a brand that took years to build. The rep sells in your voice, with your policies and your products, tuned with your team.
Honesty
Measured like a hire
Every store launches behind an A/B split against its own baseline, and we only call a result a result when it's statistically meaningful. If the rep doesn't earn its keep, the data says so.
Compounding
Every conversation compounds
Each question a shopper asks becomes first-party intent data the brand has never had — the objections, comparisons, and hesitations no analytics tool can see — and makes the next answer better.
The team
Two technical co-founders who've seen this problem from every angle — Google, Anduril, Verkada, MongoDB, Sephora, Reevo. Based in San Francisco, backed by Y Combinator.
Kratik Agrawal
Co-founder & CEO
Started on Google's ads team, going deep on how segmentation and matching work at scale. From there he trained ML models at Anduril and launched new enterprise lines at Verkada, then led Conversational Intelligence at Reevo — an $80M Series A backed by Khosla and Kleiner Perkins — building the systems that helped B2B sellers walk into every meeting better prepared. San Jose native, UCLA CS.
Varun Kandula
Co-founder & CTO
Data and infrastructure to the core. He worked on the data context team at MongoDB and built context graphs at Reevo, and consulted for Sephora on their AI features in 2023. During COVID he ran his own e-commerce store at $10K a month — so he knows firsthand what it takes to run ads and fight for conversions. D1 squash player. UC Berkeley.
The fastest way to know us is a demo
We'll run the rep against your own catalog — your products, your policies, your hardest customer questions. And if you'd rather build this than buy it, we're hiring.