When a voter asks a chatbot about a candidate, the answer often comes from Ballotpedia. Research from the Center for Campaign Innovation found it cited in about 30% of chatbot answers about candidates, and in 42% of Google Gemini’s. Eric sits down with Leslie Graves, founder and CEO of Ballotpedia, to find out why. Her answer is simple: structure, original content and plain text.

They dig into why candidates should fill out Ballotpedia’s Candidate Connection survey. A race’s page on Ballotpedia can draw page views equal to 20 to 30% of the people who vote in it. Leave gaps on your own website, and AI may fill them with a four-year-old Reddit thread about something you said at a town hall. Leslie’s advice: don’t game the algorithm. Write for people.

Ballotpedia: https://ballotpedia.org

Visit our website: CampaignTrend.com

Just 4% of voters say they used an AI tool for election information in the past month. But 38% got it from a Google AI summary, and about half of them didn’t realize it was AI. Eric sits down with B.J. Martino, president and CEO of The Tarrance Group and part of the bipartisan research team behind the AAPC Foundation’s new national survey of likely voters on artificial intelligence.

They dig into the partisan gap on whether AI will do more harm than good, why just 1% of voters call AI their most credible source of election information, and where voters draw the line on AI-generated content. B.J. also has advice for the final stretch: know what the AI summaries say about your campaign, and be ready to show your work when someone cries deepfake.

Visit our website: CampaignTrend.com

Did the internet let politics down, or did we? Eric sits down with his friend Katie Harbath, founder and CEO of Anchor Change and author of the new book Disrupting Politics: A Front Row Seat to the Collision of Technology and Democracy. Before spending a decade running global elections policy at Facebook, Katie built GOP.com, ran digital for Rudy Giuliani's presidential campaign and led digital at the NRSC, so she's watched this collision from both sides. Eric and Katie trace how real-time analytics pushed campaigns toward whatever got the most clicks, and why social media turned "villainous overnight" after the 2016 election. They also talk about the whiplash campaigns felt when the same tactics went from brilliant to sinister in just four years. Then they turn to AI, where the pendulum has swung the other way. Katie closes with two lessons from the book: good intentions are not safeguards, and panic responsibly.

Get the book: Disrupting Politics

Visit our website: CampaignTrend.com

Almost everything written about AI and elections this cycle is about what it will do to voters. Paul Allen thinks the bigger story is what it can do for them. This week, Eric sits down with Paul, co-founder of Ancestry.com and founder and CEO of SOAR and Citizen Portal, who is applying the lessons of turning a huge unread archive into something personal to the records almost no one reads: local government meetings, ordinances and regulations.

They get into how Citizen Portal transcribes public meetings and links every AI-written story back to the tape; a "newspaper business in a box" meant to make one-person local newsrooms viable again; and why AI agents may be the only way clerks keep up with a flood of public records requests. Eric pushes on whether elected officials will welcome that kind of transparency or just find new places to hide — and Paul challenges the premise.

Visit our website: CampaignTrend.com

This week, Eric Wilson is the guest. Right after Labor Day, Eric joined Prompt.io president Farz Sokhansanj on Prompt.io's Pulse Check webinar to take stock of where campaigns and technology stand heading into the midterms.

They cover the four kinds of Americans campaigns are trying to reach, and why only about 9% of them are the ones actually talking about politics. Whether 2024 really was the "podcast election," and why podcasts haven't worked down the ballot. How AI is showing up in this cycle's ads, and why it should be spice, not the whole meal. The 42% of Republicans in toss-up races who don't have an issues page on their website, even as AI search leans on those pages. Why over-targeting leaves votes on the table. What voters actually mean when they ask for a candidate who's "authentic." And why vendors need to be building case studies right now.

As Eric puts it: the winner isn't the person with the most views. It's the person with the most votes.

Visit our website: CampaignTrend.com

Eric is joined by Dustin Tropp, founder of Republican Jobs, to unpack the 2026 Political Careers Survey — 427 political professionals on pay, benefits, and whether they're sticking around.

The headline finding: it's not the job that determines whether campaign work is a real career, it's the employer. Staff at a firm or committee report just 16% with zero benefits. Staff hired directly by a campaign? 64% with nothing. Same candidates, same work, wildly different security — and Dustin explains why campaigns can't fix it on their own.

They also get into why senior campaign hires are taking two to three months to fill this cycle, why hiring is shifting from March to November, what happens once a staffer crosses the $125K mark, and why field organizers — the least automatable job in politics — are still among the worst paid. Plus: what the survey found on AI (42% say it's changed how they work, only 12% see it as a threat), and why employers aren't paying for the extra output they're getting from it yet.

Dustin also shares real-time hiring data most surveys miss: 107 open roles, where the shortages are worst, and what he's telling clients about paying below-market rates.

Visit our website: CampaignTrend.com

Can you trust a poll when no human ever picked up the phone? Eric sits down with Matthew Hanauer, co-founder of Civly, to unpack synthetic polling — building AI simulations of real, individually-modeled voters (using public voter file, demographic, and donation data) to predict how they'd respond to a survey. Matt walks through how Civly back-tests its models against real election results, why asking AI yes/no questions produces unrealistically extreme answers, why piling on more data (like social media) can actually make predictions worse, and how turnout modeling benefits from more than just "did they vote last time." Eric and Matt land on where AI polling actually fits into a campaign's toolkit in 2026: filling the gap for downballot races that could never afford a traditional poll, and pressure-testing messages before committing budget to the real thing.

Visit our website: CampaignTrend.com

Opposition research has always rewarded whoever digs deepest. AI is changing who can afford to dig — and how much stays buried. This week, Eric Wilson sits down with John Artunkal, founder and CEO of Argus AI, who came to politics through open-source intelligence, trained by Bellingcat and schooled at Oxford in the social science of the internet.

They get into how AI is — and isn't — transforming the vetting business: digital footprint mapping, facial recognition, and linguistic forensics; why self-vetting is now table stakes for down-ballot candidates; and the trade-off every research tool has to make between flagging too much and missing the one thing that ends a campaign. John also makes the case that candidates shouldn't over-sanitize themselves. Voters reward authenticity — the job is knowing the difference between a rough edge and a career-ender.

Visit our website: CampaignTrend.com