The Tunnl Story
How campaign grit became the modern moat for intelligence
The Problem We Saw
Tunnl’s story begins in 2002, when co-founder Brent Seaborn and Alex Gage were leading a new way to identify persuadable voters. Their first microtargeting pilot was in Michigan, a critical battleground state, alongside Michael Meyers, who would later join them in launching TargetPoint. Months later, Sara Fagen was dispatched from the White House to build the data and targeting operation for President George W. Bush’s re-election campaign. Still reeling from the razor-thin 2000 race, the campaign needed more than instinct, it needed a provision system for finding and persuading the voters who would decide the outcome.
The answer was TargetPoint’s early microtargeting system: a large-scale survey paired with aggregated data sets that projected issue motivation, vote intent, and turnout, not for basic demographic segments, but for reaching individual voters.
Long before cheap compute and modern AI, our founders were chasing a simple truth: when every vote matters, survey research alone isn’t enough. They began pairing large-scale surveys with early machine learning, turning research into precise, individual-level voter targets. By the November 2004 election, the shift was clear. Campaigns would never operate on intuition alone again.
Eight years later, that thinking found a new challenge. Alex Lundry, leading data for Mitt Romney’s 2012 campaign, was trying to match what Obama’s team had pioneered in 2008: more precise, more effective media targeting.
The problem wasn’t a lack of data. It was what happened after. Research and polling operated in isolation from paid media. Insights lived in decks as bulky toplines and printed crosstabs, disconnected from the people they described. There was no direct path from what you learned to who to reach, how to reach them, or how to measure impact at the individual level. The largest budget items were the least data-driven and the least accountable.
That gap wasn’t just inefficient, it was structural. Data sat across siloed tools, vendors, and teams, forcing campaigns to manually stitch together audiences, messaging, and measurement. In high-stakes environments, that doesn’t just slow you down. It costs outcomes.
So they set out to build something different. That work became Deep Root Analytics.
Building the Foundation: Microtargeting
In 2003, TargetPoint was founded alongside a pivotal presidential race to solve this problem head-on. At a time before “AI” was a mainstream term, these teams were already building and deploying machine learning systems to power microtargeting. They weren’t just surveying populations, they were training models to predict behavior, linking attitudes to actions and actions to outcomes at scale.
For nearly a decade, TargetPoint operated at the frontier of political data science, pioneering predictive modeling approaches that laid the groundwork for modern AI-driven communications. By combining large-scale data modeling, opinion research, and audience identification, the firm helped campaigns and advocacy organizations move beyond static insights toward probabilistic, forward-looking strategies—reaching the right voters with a level of precision that was, at the time, unprecedented.
Extending Precision to Media: Deep Root Analytics
Building on the foundation established at TargetPoint, Deep Root extended predictive intelligence into media by pairing early machine learning models with emerging datasets like set-top box and later Automatic Content Recognition (ACR) data. For the first time, campaigns could apply the same precision to how and where they reached those same people.
With nearly 80% of media spend flowing into linear TV at the time, this shift mattered. Audiences were no longer bought as broad demographics, but as defined segments. And each had to be reached with tailored messaging on the programs they were most likely to watch. The era of proxy-based media buying was coming to an end.
But the real breakthrough wasn’t just better targeting, it was operationalizing that intelligence at scale. Where TargetPoint modeled behavior, Deep Root embedded those models directly into planning, activation, and measurement across TV and digital. Campaigns could move from insight to execution, and adjust in real time while budgets were still in flight.
Together, the two efforts created infrastructure well ahead of its time: individual-level identity resolution at national scale, continuous model-driven optimization, and measurement tied to real-world behavior change across channels.
In 2020, Deep Root brought it full circle: it acquired key intellectual property and microtargeting capabilities from TargetPoint. A decade of research and activation capabilities were finally unified, bringing prediction and execution into a single, more connected foundation.
Building for a Broader Market: Tunnl Launches
In 2021, that expertise formally converged into Tunnl, a new vision built on nearly two decades of machine learning, training data, and real-world campaign execution. By unifying the methodologies of TargetPoint and Deep Root into a production-ready, closed-loop system, Tunnl transformed years of predictive modeling into a scalable intelligence platform.
The result is more than just an AI product. It’s over a decade of infrastructure and 20+ years of domain experience that most recent AI companies simply don’t have, enabling Tunnl to move from insight to real-world impact in ways others cannot.
While a new wave of AI companies emerged focused on generating answers, Tunnl was built on a different premise: prediction is only valuable if it drives action. From day one, the platform was designed to connect intelligence directly to execution and measured outcomes.
Where campaigns had always operated under relentless pressure—speed, precision, accountability—corporate communicators, advocacy organizations, and brands were beginning to face the same conditions. Tunnl was built to meet that moment, bringing AI-driven decisioning out of political campaigns and into the broader market.
To serve that market with full focus, Tunnl was established as an independent company outside of candidate campaigns: purpose-built for brands, associations, advocacy organizations, and agencies navigating complex environments.
Following its launch and initial funding in 2021, the Tunnl Platform debuted in 2022 with Audience Explorer and TV planning tools. But these were not point solutions, they were early surfaces on top of a deeper system: identity resolution, predictive modeling, and activation infrastructure designed to move from insight to influence.
Expanding the Platform
As the platform matured, Tunnl expanded its capabilities to support organizations where influence, not just awareness, drives outcomes. In 2023, the Premier Audience Suite extended predictive targeting into high-stakes sectors like healthcare, finance, and public policy. Tunnl was identifying not just audiences, but the specific decision-makers within them.
In 2024, Tunnl launched Reach & Frequency, bringing modern, model-driven measurement to linear television. This bridged a critical gap: applying advanced audience intelligence to one of the largest and least disrupted media channels.
Completing the Vision: From AI to Outcomes
2025 marked a major evolution. Tunnl delivered a fully integrated system by enhancing Audience Explorer, launching Campaign Effectiveness, and connecting exposure directly to measurable shifts in sentiment and behavior. For the first time, organizations could see, not assume, whether their campaigns were working, regardless of channel.
That same year, Tunnl unified its stakeholder targeting capabilities under Halo, a system designed to identify, prioritize, and influence the most critical decision-makers across policy, business, and public discourse and the spheres around them.
The integration of TargetPoint Consulting in 2025 completed a 20+ year arc. Research, modeling, and activation were no longer adjacent capabilities: they became a single, continuous intelligence system. The models didn’t just inform strategy; they powered it in real time.
Tunnl also introduced generative AI—but in a fundamentally different way than LLM-native platforms. Rather than scraping the internet or generating surface-level insights, Tunnl is applying AI directly to its proprietary, structured, longitudinal data moat. Beyond aggregating critical data sources, the platform helps users interpret complex signals and make faster decisions, all within this closed system.
Because at Tunnl, the belief is simple: understanding is cheap. Changing outcomes is hard.
What’s Next & Why It Matters
In 2026, Tunnl takes the next step with the launch of Research Studio. Bringing research results into the platform and connecting them to the modeling, audience building, activation, and measurement was by design. This launch completed the feedback loop: every interaction informs the next decision, and every campaign makes the system smarter.
This is not an AI tool layered on top of workflows. It is an AI-powered system of record for influence.
While LLM-native companies promise instant understanding, Tunnl delivers something far more valuable: the ability to predict, reach, and change the audiences that matter—continuously, measurably, and at scale.
The same pressures that once defined political campaigns now define every industry. Speed, precision, and accountability are no longer advantages—they are requirements.
Tunnl was built by teams who understand what failure costs in those environments. That is why organizations using Tunnl don’t just generate insights.
They drive outcomes—and get smarter every day they run.