Jason Damata: What is Newton Research?
John Hoctor: Newton Research has built a platform called Newton, a team of highly specialized AI agents that can be embedded into every stage of the media workflow—from planning and targeting to buying, optimization, and reporting. Unlike generic chatbots or open-ended co-pilots, these agents are trained specifically on the workflows, datasets, and decision patterns unique to media and marketing.
Jason Damata: How did this idea come about?
John Hoctor: It comes from decades in the industry. My co-founder Matt and I have worked together for 26 years, most recently at Data Plus Math, where we focused on measurement and reporting across linear and CTV. Over the years, we saw every company wishing they had more analysts and data scientists. With AI, we saw the chance to make sophisticated measurement more accessible—helping teams handle the exploding complexity of data, expectations, and campaign dynamics.
Jason Damata: With so much investment in AI, how do you differentiate Newton?
John Hoctor: In the age of AI, moats are being redefined. Many general-purpose tools like Gemini, ChatGPT, or Claude are great at broad tasks but struggle with industry-specific analytics. Ask Claude to build a media mix model and it may return brittle, impractical code. Our moat is specialization. Newton agents are trained to reflect real-world campaign dynamics and produce analytics that marketers can actually use.
Jason Damata: How do you build that specialization into Newton?
John Hoctor: We’re constantly teaching Newton new skills and methodologies—like giving it a “media analytics handbook” or a “master’s in marketing science.” That means when you ask Newton to set up an incrementality test, it understands the request, has examples to pull from, and knows how to apply the right methodology—unlike a generic chatbot that’s been trained on the open web.
Jason Damata: How do you protect customer data?
John Hoctor: Newton was designed for large brands, publishers, and agencies that can’t risk sharing proprietary knowledge. Newton runs as a containerized application where the data lives—there’s zero data movement outside the customer’s environment. Customers can also teach Newton their own methods, scripts, or notebooks. That training stays entirely within their local environment; it never goes back into our algorithms or knowledge base.
Jason Damata: Does that make Newton sticky from a business perspective?
John Hoctor: Exactly. Customers aren’t handing over data—they’re training an extension of their team that never takes vacations or leaves. Importantly, Newton doesn’t replace people; it makes overworked analytics teams more productive. They’ve often tried using ChatGPT or Claude, but those tools don’t behave the way they need. With Newton, the agents already “think like they think” and can be trained further, enabling teams to infuse data and analytics into much more of the workflow.
Jason Damata: How do you see Newton impacting television?
John Hoctor: Measurement tells you what happened in the past. Newton agents enable predictive media modeling, where analytics can be applied in real time to shape what happens next. For TV, that means campaigns can be measured while in flight, with insights feeding directly into optimization for the next decision. Customers are already using Newton this way, shifting from retrospective reporting to active, predictive decision-making.
Jason Damata: Do you see a “battle of the bots” ahead in media optimization?
John Hoctor: I think it’s more likely a cooperation of bots. Emerging protocols like MCP and agent-to-agent could allow intelligent agents to interoperate. If implemented wisely, our industry has the chance to build an ecosystem that works for everyone—without repeating the mistakes of monopolies and walled gardens.