Dvir Ginzburg's Speaker Profile
Every Enterprise Has Revenue Locked Inside Its Customer Conversations.
Dvir Ginzburg Built
the Patents That Unlock It.
Meet Dvir Ginzburg — inventor of patented Interaction Mining technology and the founder defining what revenue-generating agentic AI means for enterprises.
Trusted by
The One-Liner
(for host intros)
Dvir Ginzburg holds two patents for Interaction Mining — the process that distills what top performers do in customer conversations and deploys it as enterprise-grade AI agents that generate revenue at scale. He founded Insait in 2022 because he saw that every enterprise already had its answer to AI: its best people. The problem was that those people couldn't scale. His work proves they now can.
Biography
Who is Dvir Ginzburg?
Dvir Ginzburg started as an academic — Ph.D. in Geometric Deep Learning under one of Israel’s leading deep-learning advisors — then spent years inside Microsoft as a recommendation-systems researcher. The technical foundation he built there runs inside Insait today.
What he saw at scale was a gap no AI vendor was closing. Every enterprise has a ceiling: the number of top performers it can hire. On a contact center floor, a lending desk, a wealth advisory team, an insurance sales operation — the pattern repeats everywhere. A small group of people convert at multiples of everyone else. The institution knows it. It tries to train for it. It never fully replicates it.
In 2022, Dvir founded Insait to close that gap. Not with a better chatbot. Not by improving deflection rates. By going upstream: taking the raw interaction data — the calls, the chats, the live conversations — distilling exactly what makes top performers exceptional, and deploying it
as enterprise-grade AI agents that capture revenue at scale. That process is Interaction Mining. Dvir holds two granted patents on it.
Three years in, Insait runs in production across Tier-1 institutions in the US, Israel, EU, and APAC — deployed wherever high-stakes customer conversations happen and revenue is on the line.
His thesis hasn’t changed since day one: the answer to AI isn’t a better deflection tool. It’s an AI
that earns its place on the revenue line.
Resources
Highlights
- CEO & Co-Founder of Insait: Series A enterprise AI company running in production across Tier-1 institutions in the US, Israel, EU, and APAC
- Inventor of Interaction Mining two granted patents on the process that distills top- performer behavior from raw interaction data into enterprise-grade AI agents
- Ph.D. in Geometric Deep Learning with academic research that directly shaped Insait's patented hybrid recommendation-engine + LLM architecture
- Ex-Microsoft recommendation-systems researcher the technical depth to build defensible IP, not wrap a third-party model
- Built Insait into a global operation Customers include iLending, Bank Leumi, FIBI, Harel Insurance, PassportCard, OnDeck
Book Dvir Ginzburg now!
Media appearances
Listen to Dvir Ginzburg's previous interviews
Topics
What your listeners can learn
Topic #1
Revenue Per Interaction Is the Only AI Metric That Will Survive 2026 — And Most
Enterprises Are Still Measuring the Wrong Thing
AI vendors spent five years selling deflection: cost per call down, tickets resolved, support offloaded. The enterprises pulling ahead now are measuring something different: revenue per interaction. Dvir walks through what changes when you flip the KPI — what gets built, what gets measured, what gets retired — and why the shift is irreversible across any industry where a customer conversation can end in a sale.
Listener takeaways:
- Why “cost per call” is a 2018 metric and “revenue per interaction” is the 2026 one
- The single question that reveals whether any AI vendor is built for deflection or conversion
- What Interaction Mining found in the data: the performance gap between top reps and average reps is larger than most executives think — and it is fully measurable
- What production deployments actually look like: real results, no consultant speak
Topic #2
Interaction Mining: The Patented Process That Turns Your Best Customer Conversations Into Enterprise-Grade AI
Every AI vendor walks into an enterprise and asks for a knowledge base, a playbook, an FAQ document. Insait asks for the conversations. Dvir explains what Interaction Mining actually does — how it takes raw voice, chat, and live interaction data; what it distills from that data; and why what comes out the other side is categorically different from a configured AI agent. Two patents’ worth of different.
Listener takeaways:
- What Interaction Mining is and why it required two patents to protect
- The difference between an AI agent configured to a template and one compiled from your actual interaction reality
- Why the data advantage compounds: every interaction captured sharpens every agent that runs after it
- What compliance-grade AI means in practice — and why it is the unlock for enterprise buyers, not a legal checkbox
Topic #3
Your Enterprise Already Has the Answer to AI. It’s Your Top Performers.
The most defensible AI strategy an enterprise can run has nothing to do with which model it picks. It has everything to do with which interactions it learns from. Dvir makes the case that every company has already mapped its performance ceiling in its conversation data — and that Interaction Mining is the process that raises that ceiling. Not by adding headcount. By deploying what your best people already know, at a scale they cannot reach on their own.
Listener takeaways:
- Why top performers are the most underutilized asset in any enterprise
- What it means to build AI from your reality, not configure it to a vendor’s template
- The data flywheel: how every interaction captured sharpens every agent that follows
- Why this applies across any enterprise where customer conversations generate revenue — not only financial services
Topic #4
Your Suggestion…
We’ll be happy to check with the team and confirm whether it’s something they can cover in depth and deliver value to your audience.