Higher Ground Labs LLC

10/25/2026 | Press release | Distributed by Public on 10/26/2026 08:55

Meet Haystack: Social Listening Built for Democrats

Introducing Higher Ground Labs' newest Agentic AI investment:

"We want Democratic campaigns to know what's happening today and drive the news cycle tomorrow," says Nathan Wyeth, co-founder of Haystack, "rather than finding out three days later when it shows up in their feeds."

In a media environment where everyone publishes, and every audience sees a different feed, Haystack gives political communicators AI-native social media intelligence, built by operators who have done the work themselves. Haystack is one of five companies Higher Ground Labs invested in through our 2026 Agentic AI Open Call.

A Higher Ground Labs origin story

Liz Utrup brings 20 years of experience in political media and messaging research. As a 2025 Higher Ground Institute Media Innovation Fellow, she led the social listening study that inspired Haystack. Nathan Wyeth spent a decade in product and commercial leadership at Sunrun and Octopus Energy. He has worked in Democratic politics since 2004 and co-founded Clean Energy for Biden.

The two were introduced by HGL's own Betsy Hoover. Liz had been developing a case for tech-driven social listening as a core part of Democratic research. Nathan had independently arrived at a strikingly similar idea, and when he brought it to Betsy, she put them on an email chain with a simple message: you two need to talk. In what Nathan calls "pure serendipity," they were on similar timelines and ready to start a new chapter.

For Liz, the mission goes back to 2008. "When we did Obama '08, we had what was called 'new media' at the time," she says. "We were so engaged and innovative with what technology could do for a political operation, and really leading the way. And we have totally lost the lead."

Haystack became their answer: a way to give Democratic communicators the visibility into today's online media environment they've been missing.

"It's like fencing off the portion of the internet that is most relevant to you and your purpose."

- Liz Utrup, Co-Founder

How Haystack works

  1. Organize a feed. Tell Haystack what you want to track, and it finds the relevant accounts, forums, and hashtags. It brings them into one feed across platforms, including public accounts of any size and the comments beneath posts. It can even infer location from clues buried in content. If a creator mentions attending a Cowboys game at AT&T Stadium, Haystack infers that the account might be based in Dallas, Texas.
  2. Find the needles. Tell Haystack the kinds of posts you don't want to miss, and it reviews every one: text, images, and video transcripts. It isn't keyword search. Haystack looks for meaning and context, so a team can ask for something like "people talking about challenges they've personally experienced accessing health care."
  3. Understand what happened. Ask Haystack questions about your feed, and it delivers answers to your inbox daily or as in-depth research. It tracks trends, spiking topics, and inflection points, measures shifts in favorability and emotion toward people, organizations, places, or issues, and runs deep-dive research across thousands of posts in minutes.

"You have a short window to act in an impactful way on social channels," Liz explains, "and this puts you in a place where you can take action rather than do the tedious work to figure out what's happening across the internet."

Haystack in action

Since June, Haystack has powered a team at a leading national think tank as it sources and amplifies original social media content affirming progressive values. In that time, Haystack has:

  • sourced roughly 1 million publicly available accounts, analyzed hundreds of thousands, and identified 10,000 accounts of active creators connected to priority communities and identities;
  • ingested and reviewed more than 1 million posts and transcribed 10,000 hours of video;
  • automatically surfaced message-aligned content, especially first-person, emotionally driven short-form video.

Haystack addressed an urgent need for a sophisticated team that knew exactly what existing tools couldn't do. "They had a sense from the beginning of the tools they had access to already and what those couldn't do," Nathan says. "From our first conversations, it was clear that we were filling a gap."

Built for this moment

Haystack went from concept in January to a fully functioning product in mid-June. "AI is empowering operatives like us to build elaborate, sophisticated tools for our greatest challenges," Liz says. "It's so important that we leverage this moment to use AI for good."

Nathan argues the timing is no coincidence. He estimates the volume of content comms teams face has grown by roughly 50x in two decades, far more than any person or team could ever take in. "If there's one thing LLMs are good at, it's reading a lot of text and summarizing it, or listening to a lot of video and summarizing it," he says. "We've been able to step into this at a moment when the technology has met the need."

What's next

Liz and Nathan are now focused on scaling Haystack's user base, enabling users to build programs based on the in-depth, real-time knowledge the app enables, and helping the Democratic ecosystem invest more deeply in organic media. Their goal is to make social listening a core pillar of campaign and advocacy strategy, alongside organizing, paid, and earned media, rather than an afterthought.

"There's been a smart, concerted effort to grow the amount of progressive content being pushed out. We want to empower and enable that effort by making organic strategies faster and easier to execute."

- Nathan Wyeth, Co-Founder

Ready to get started? Reach the Haystack team at [email protected].

SEE HAYSTACK

Haystack is one of five companies Higher Ground Labs invested in through our 2026 Agentic AI Open Call.

Higher Ground Labs LLC published this content on October 25, 2026, and is solely responsible for the information contained herein. Distributed via Public Technologies (PUBT), unedited and unaltered, on October 26, 2026 at 14:55 UTC. If you believe the information included in the content is inaccurate or outdated and requires editing or removal, please contact us at [email protected]