Tunnl, an audience intelligence company based in Arlington, Virginia, today said it has partnered with DeepIntent to put its opinion-research segments inside DeepIntent's healthcare programmatic platform, according to Tunnl. The company says it supplied 131 audiences to start, activated on a data CPM billed on top of media. No case study with DeepIntent exists yet.

In Short

Tunnl, which turns opinion polling into groups of people that advertisers can target, today said 131 of its groups are now available to buy through DeepIntent, an ad-buying platform for drugmakers and health brands. Health advertisers usually sort people by age or illness, so groups built on things like reluctance to seek care or trust in FDA drug approvals give them a different way to decide who sees an ad. Tunnl says each group carries an extra per-thousand-impressions data fee, and with no results from the DeepIntent pairing yet, how well the groups perform is not public.

What the companies describe

Tunnl, which calls itself an AI-powered audience intelligence platform, said in a release today that marketers can reach its research-backed audience segments through DeepIntent's programmatic setup for health advertisers, according to Tunnl. The integration, in the release's wording, brings differentiated audience solutions to a category where "privacy expectations, regulatory requirements, and sensitive health decisions make patient communication especially complex."

The segments draw on recent opinion research. According to Tunnl, that research adds four dimensions to campaign planning for marketers who have organised audiences mainly by demographics or condition: hesitancy toward seeking care, attitudes toward risk, openness to digital healthcare services, and the factors that encourage engagement. The release illustrates the idea with three patients who share a demographic profile and a healthcare need. One is hesitant about seeking care, another needs caregiver support to navigate the system, a third prefers digital services to save time. Demographic and condition-based categories alone, Tunnl says, do not explain those differences.

Sara Fagen, chief executive and co-founder of Tunnl, put the argument this way: "Demographics and conditions provide part of the picture, but they don't fully explain what makes someone hesitant to seek care, how they think about risk, or whether they're comfortable engaging with healthcare services digitally." According to DeepIntent, marketers are turning to partners with access to a wide range of audience solutions as the advertising environment grows more complex. Natalie Mancuso, senior vice president of data partnerships at DeepIntent, said the company is "bringing new dimensions to how healthcare and pharma marketers better understand the audiences they are trying to reach."

Tunnl's own description helps place it. According to the company, it was founded in 2021 by Fagen and Brent Seaborn, whom it describes as data and political strategy veterans, and it serves Fortune 500 brands, media agencies, trade associations and advocacy organisations. It says its platform connects survey research, insights, audience targeting, activation and campaign measurement in a single closed-loop product. Healthcare, the release adds, is "just one application": the same modelling approach is said to support consumer, advocacy, media and political audiences, and Tunnl expects to add further activation partnerships. It names none.

131 audiences, priced on top of media

The release itself gave no catalogue size or price. Tunnl supplied both afterwards. According to Tunnl, the company has provided 131 audiences to start, available off the shelf in DeepIntent's Audience Marketplace and activatable in DeepIntent's healthcare DSP. The examples it cites are COVID vaccine intent audiences, flu shot intent audiences, birth control funding support audiences, childhood vaccination audiences and confidence in FDA medication approval audiences. Tunnl describes all of them as behaviour and perception based.

On cost, Tunnl says the arrangement works much like other audience intelligence and DSP partnerships. Activation is based on a data CPM, billed on top of media and applied to every thousand impressions served to that segment. Tunnl says the specific marketplace pricing is not public yet. A data fee of that kind sits outside the media price a buyer negotiates with the platform, so the all-in cost of reaching a segment depends on a rate that neither company has disclosed.

How the segments are built

Tunnl's account of its method is brief. According to Tunnl, the audiences rest on stated opinions and behaviours collected in large-sample surveys. The survey questions inform models that apply to the entire US population, and the output is a set of predictive scores used to create segments. Because the models score everyone, most of the people placed in a segment were never asked the questions, and their attitudes are inferred rather than reported.

Tunnl argues the approach suits health advertising. In areas such as healthcare, where privacy is critical, the company says perceptions and behaviours can make for a better vehicle for engagement than condition-specific data or other highly personal information. All of its data, according to Tunnl, connects back to a proprietary US identity graph housed in a secure environment that follows all federal and state privacy compliance for consumer data. That is a statement by the vendor, and the companies have not published an audit or certification.

What the account leaves out is substantial. Neither company states the survey sample size, how often the research is refreshed, how the predictive scores were validated, how accurate they are at the individual level, or how many people the identity graph covers. Nor is it clear how a score reaches an individual device or record once a segment goes to a bidder. The scope is the United States, since the models and the graph are described in US terms.

Where the deal sits in DeepIntent's stack

DeepIntent describes itself as a healthcare demand-side platform, and PPC Land's coverage over the past 14 months shows a company that has mostly been adding places to buy and tools to plan. In August 2025 it rolled out a streaming package for free ad-supported channels. On March 19, 2026 it opened Helix, a marketing cloud covering 3.7 million providers and 240 million patient lives, figures reported in coverage of its later agentic planning tool.

Late May brought a burst of supply. On May 26, 2026, epocrates opened its clinician workflow to programmatic buying through DeepIntent, giving access to more than one million active clinicians. Within days, three supply deals with OptimizeRx, epocrates and Lane4.io brought point-of-care placements into the same buying interface, spanning six electronic health record systems and 158,000 doctors. June 9 added verified live sports inventory, and an August integration with Vistar Media extended the reach to digital out-of-home screens.

Tooling followed. On September 10, 2026 DeepIntent opened Cora, a conversational planning layer, as an open beta to all clients, with a claimed 50% cut in planning time. The same coverage describes a push into market research, forecasting and linear television, a long way from the auction pipes where the company started. A partnership with a company whose product starts from opinion research sits comfortably inside that direction of travel, though the release draws no such line.

The Tunnl arrangement differs in kind from most of those moves. Supply additions widen where an advertisement can run. A data partnership changes the question of who is considered worth reaching, and on what basis. The release does not say whether the segments live inside Helix, in the standard bidding environment, or both, though Tunnl places them in the Audience Marketplace and the healthcare DSP.

Performance claims and regulation

No results exist yet for the pairing. According to Tunnl, the DeepIntent integration has only just been finalised, so there is no case study with DeepIntent. The company points instead to other healthcare work with major healthcare groups, medical device manufacturers and biotech organisations, with results it describes as ranging from increased brand awareness to a 70% reduction in cost per click. Those are vendor-supplied figures. Tunnl names no client, period, baseline or sample, and a cost-per-click result says little about the outcomes pharma budgets are normally graded on.

DeepIntent's description of itself says its platform delivers measurable audience quality and script lift at scale. Script lift, the change in prescription volume that a campaign can be shown to have produced, is the currency in which pharma budgets tend to be graded, and such studies are observational rather than experimental, so they estimate association, not causation. Neither the release nor the details Tunnl supplied offer a script lift figure for the segments. The strongest effectiveness language comes from Mancuso, who said the segments enable clients "to build more effective and relevant campaigns." Tunnl's own framing is more modest: more ways to understand decisions, more options for education and service awareness.

Regulation gets similarly light treatment. DeepIntent's boilerplate calls its marketing "privacy-safe", but the release does not mention HIPAA, the 1996 US law governing protected health information. The statute has no certification scheme, so the phrase "HIPAA-compliant" describes a self-assessment rather than an audit, and a recent DeepIntent release for Cora likewise dropped the word in favour of privacy-safe framing without explanation. Several segment names refer directly to health decisions, among them vaccine intent and birth control funding support. How such modelled scores are classified under health privacy rules and platform policies is a question the companies have not addressed.

Scrutiny of the keys that link health-related data to people is also rising. Italy's data protection authority fined IQVIA 7 million euros over the records of roughly one million patients that the company had described as anonymous, in a decision adopted on September 23, 2026 and published on October 2. The regulator's reasoning was that a hash built to preserve uniqueness does not make a record anonymous. Tunnl's segments rest on opinion-research modelling rather than clinical records, so the ruling is context rather than comparison. It does show the sort of question a health data partnership can attract.

Why it matters to marketers

DeepIntent describes its platform as uniting media, identity and real-world clinical data, signals that record what has happened to a person. Attitudinal research tries to capture something else: how a person thinks about care, risk and digital services. The Helix figures above give a sense of the clinical side's scale, and Tunnl supplies the other kind. Whether the combination improves outcomes is the open question, since the companies assert added dimensions, not added performance.

There is also a cross-category angle. Tunnl lists politics among its categories, and the partnership lands less than four weeks before the US midterm elections on November 3, 2026, in a cycle that AdImpact projects at $11.6bn of political advertising. The release does not say whether the healthcare segments and the political ones share research panels or modelling, and it gives no sign that they do.

Cost is the one item that is now partly answered: a data CPM on top of media, at a rate yet to be published. The open items are the rate itself, the survey scale behind the scores, their freshness and accuracy, and measured lift. Until DeepIntent or Tunnl publishes figures, or a pharma advertiser reports a study, the partnership adds an option to the menu with no public evidence of what it delivers.

Timeline

  • 2021 - Tunnl is founded by Sara Fagen and Brent Seaborn, according to Tunnl.
  • August 21, 2025 - DeepIntent debuts its HealthFirst FAST Package for streaming advertising.
  • March 19, 2026 - DeepIntent opens Helix, a marketing cloud covering 3.7 million providers and 240 million patient lives.
  • May 26, 2026 - epocrates opens programmatic access to more than one million active clinicians via DeepIntent.
  • Late May 2026 - DeepIntent closes three point-of-care supply deals with OptimizeRx, epocrates and Lane4.io.
  • June 9, 2026 - DeepIntent adds verified live sports and events inventory for pharma buyers.
  • June 15, 2026 - DeepIntent rolls out Helix AI, with beta results showing a 50% cut in planning time.
  • August 2026 - Vistar Media's point-of-care and digital out-of-home inventory enters DeepIntent's platform.
  • September 10, 2026 - DeepIntent opens Cora to all clients as an open beta.
  • September 23, 2026 - Italy's data protection authority adopts its decision fining IQVIA 7 million euros; it is published on October 2.
  • Today - Tunnl and DeepIntent partnership release, with 131 audiences available to start, according to Tunnl.

Summary

  • Who: Tunnl, an AI-powered audience intelligence company in Arlington, Virginia, founded in 2021 by Sara Fagen and Brent Seaborn, and DeepIntent, a healthcare marketing platform. Fagen is Tunnl's chief executive and co-founder; Natalie Mancuso is DeepIntent's senior vice president of data partnerships.
  • What: A partnership making Tunnl's research-backed audience segments, defined by hesitancy toward care, risk attitudes, digital openness and engagement drivers, available through DeepIntent's healthcare programmatic platform. According to Tunnl, 131 audiences are available to start, activated on a data CPM billed on top of media, built from large-sample survey models applied to the US population.
  • When: Today. The release gives no availability date, and marketplace pricing is not yet public, according to Tunnl.
  • Where: The release is datelined Arlington, Virginia. The models and identity graph are described in US terms, and the companies name no other markets.
  • Why: According to the companies, demographic and condition-based categories do not explain why similar patients act differently, and research-backed segmentation adds attitudes, motivations and engagement preferences to audience planning. For DeepIntent it widens the range of audience solutions; for Tunnl it adds an activation partner beyond its other categories. No DeepIntent case study exists yet.