August 26, 2026 · 46 min · Episode 775
The Math of Marketing: George Khachatryan on Reinforcement Learning and AI Personalization
About this episode
Anika sat down with George Khachatryan to explore the evolving landscape of AI-driven marketing, personalization, and the massive trust gap currently sitting between brands and consumers. The conversation revealed a counterintuitive truth: while 93% of marketing leaders believe AI helps them understand their customers, only 53% of consumers agree. George’s unique journey—from earning a PhD in mathematics at Cornell and building an early ed-tech AI company to founding OfferFit and navigating its $325M acquisition by Braze—offered a practical look at how reinforcement learning and predictive data science are permanently replacing the manual grind of traditional A/B testing.
In This Episode
- How a background in mathematics and early intelligent tutoring systems laid the groundwork for complex AI architectures.
- Stepping away from specialized tech to learn foundational company building through management consulting and operational transformations.
- Moving beyond rigid "Next Best Action" rules to autonomous reinforcement learning agents that experiment at the individual customer level.
- The strategic decision behind merging OfferFit with Braze in a $325M deal and why deep data science and engineering alignment matters.
- Why hidden bots, lack of transparency, and mismatched brand expectations are eroding consumer confidence.
- Blending predictive machine learning models (for timing and offers) with LLM agentic copy generation for maximum impact.
- Why students and early-career marketers must embrace AI literacy to stay competitive in a rapidly shifting job market.
Timestamps
- 00:00 Introduction: The math, the $325M exit, and the AI trust gap
- 02:00 From Cornell mathematics and early ed-tech to management consulting
- 04:18 The limitations of traditional "Next Best Action" and the birth of reinforcement learning
- 09:22 Real-world personalization: How brands like Yum Brands optimize customer engagement 12:39 The acquisition journey: Why OfferFit chose to integrate with Braze
- 16:53 The 93% vs. 53% problem: Why consumers feel misunderstood by brands
- 20:41 Hybrid AI systems: Stitching together LLM copywriting and statistical decisioning
- 23:33 The personal cost of a "Sydney is just a bot" moment: The importance of transparency
- 28:32 The reality of enterprise caution vs. consumer expectations
- 29:56 Agentic commerce and the future of bot-to-bot interactions
- 34:39 Managing negotiation bots and complex consumer retention scenarios
- 37:25 Generational divides: Gen Z's dual relationship with AI adoption and career anxiety
- 40:59 Lessons for founders: Aligning philosophical visions before a merger
- 43:35 Final thoughts: Embracing the historic and fascinating shift in marketing technology
Key Insights & Takeaways
Insight 1: Personalization Requires True One-to-One Experimentation
Traditional marketing relies on segmenting audiences and applying rigid, rule-based logic. True optimization requires reinforcement learning agents that operate at the individual customer level, autonomously testing variables like messaging, timing, and offers to discover what drives incremental engagement.
Insight 2: The Danger of Hidden Bots and Broken Trust
Consumers do not mind interacting with AI or automated tools, but they expect transparency. When a brand masks an AI agent as a human—and fails to take responsibility for its actions—it creates a deep sense of betrayal that permanently damages customer loyalty.
Insight 3: The Power of Hybrid AI Infrastructure
The most effective marketing systems do not rely on a single flavor of AI. Combining statistical reinforcement learning (to determine when and what to communicate) with LLM agentic text generation (to personalize how it is phrased) yields performance lifts of over 100%.
Insight 4: Shared Vision is the Anchor of Any Successful Acquisition
For founders navigating a potential exit, technical integration is only half the battle. Long-term success relies on frank, open discussions about philosophical alignment, shared visions, and mutual risk mitigation before signing a deal.
Insight 5: Enterprise Caution vs. Consumer Speed
While marketers are often bogged down by internal alignment and slow decision-making processes, consumers are interacting with cutting-edge AI tools daily. Brands that fail to bridge this expectation gap risk becoming obsolete.
Resources & Links Mentioned
Braze
2026 Global Customer Engagement Review
OfferFit
Yum Brands & Home Security (Client case studies discussed in the episode)
About George Khachatryan
George Khachatryan is a mathematician, entrepreneur, and executive leader serving as the Head of OfferFit by Braze. Holding a PhD in mathematics from Cornell University, George previously co-founded the ed-tech pioneer Reasoning Mind and worked as an associate partner at McKinsey. He co-founded OfferFit, an AI decisioning engine that utilizes reinforcement learning to automate marketing experimentation, culminating in its acquisition by Braze in 2025 for $325 million.
Connect with George
LinkedIn: https://www.linkedin.com/in/george-khachatryan/
Website: https://www.braze.com/
YBA was named the top PR podcast in the US by FeedSpot! Feedspot is an RSS Reader that lets you follow your favorite podcasts alongside blogs and news in one place.
Like the show? Leave us a rating or review: https://lovethepodcast.com/67940257010b317cdaa9d857
Send us a Message: https://podcastfeedback.com/67940257010b317cdaa9d857
Check out our Website: https://www.yourbrandamplified.com
Speak to my Delphi Clone: https://www.delphi.ai/amplifywithanika
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
Article
George Khachatryan of OfferFit on Reinforcement Learning, AI Personalization and the Consumer Trust Gap
George Khachatryan holds a PhD in mathematics from Cornell and co-founded OfferFit, an AI marketing company that uses reinforcement learning to personalize marketing for each customer and later merged with Braze. In his conversation with Anika Jackson, George Khachatryan explains how reinforcement learning replaces segmentation, why consumers trust AI less than marketers think, and what founders should check before agreeing to an acquisition.
How did George Khachatryan go from mathematics to marketing?
Through education technology and consulting. After his Cornell PhD, George Khachatryan built Reasoning Mind, which brought intelligent tutoring systems to more than a million students by applying mathematics to personalized instruction. He then worked in management consulting before co-founding OfferFit. The common thread is turning complex data into outcomes that serve individual people.
How does reinforcement learning change marketing personalization?
It replaces segments with decisions made for each individual customer. Traditional personalization sorts customers into groups and treats everyone in a group the same way. George Khachatryan and his co-founder, Victor Kostook, built OfferFit around reinforcement learning agents that experiment and learn on their own, adjusting the timing, tone and content of marketing for each person rather than predicting behavior and applying fixed rules.
Why don't consumers trust AI as much as marketers do?
Because brands and consumers see AI very differently, and hidden AI use damages trust. Ninety-three percent of marketing leaders believe AI helps them understand customers; only fifty-three percent of consumers agree. George Khachatryan attributes the gap less to AI's limits than to a mismatch between cautious brands and consumers whose expectations have been shaped by tools like ChatGPT. He shares his own experience with a deceptive apartment rental bot: finding out afterward that he had been talking to AI did more harm than disclosure would have.
Many companies, he argues, are too cautious with their data and leave customers underserved, while also failing to disclose AI when they do use it. His answer is to use AI and be open about it. Companies that disclose their bots and agents while delivering real personalization will stand out.
What is agentic commerce, and what should founders know about acquisitions?
Agentic commerce is bot-to-bot interaction, where consumers send AI agents to negotiate on their behalf and brands deploy algorithms to optimize offers in response. George Khachatryan cites predictions that bot-to-bot commerce will grow from nineteen percent to forty-six percent by 2026. Companies will need to redesign infrastructure and decision systems for it; the strongest will combine LLM-powered conversation with data-driven decision engines that can tell a real retention risk from a routine negotiation tactic.
On acquisitions, his advice is unsentimental: confirm that you are philosophically and strategically aligned with the acquirer before merging. Skip the honeymoon phase and run rigorous joint risk assessments. Identify where the companies might diverge, agree on how to handle it, and make sure commitment to a shared vision rests on understanding rather than optimism. OfferFit's merger with Braze worked, he says, because both sides did that vetting.
Key takeaways
- Reinforcement learning lets marketing adapt timing, tone and content for each customer instead of each segment.
- 93 percent of marketing leaders think AI helps them understand customers; 53 percent of consumers agree.
- Hidden AI use damages trust more than disclosure does.
- Brands should prepare for consumers sending AI agents to negotiate for them.
- Before an acquisition, run a joint risk assessment and confirm strategic alignment.
Listen to Anika Jackson's full conversation with George Khachatryan on Your Brand Amplified.
Get booked on podcasts
We book you on 5 to 10 high-quality podcasts every month.
Podcast Guest BookingStart your own show
Coaching and a peer network for experts building a branded podcast.
Your Branded Podcast™