<?xml version="1.0" encoding="UTF-8"?>
<!-- generated by blog/tools/build.mjs; edit the .md source -->
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Vector Political Blog</title>
    <link>https://vectorpolitical.com/blog/</link>
    <atom:link href="https://vectorpolitical.com/blog/feed.xml" rel="self" type="application/rss+xml" />
    <description>Notes on what millions of conversations with voters and donors are teaching us.</description>
    <language>en-us</language>
    <lastBuildDate>Wed, 11 Nov 2026 12:00:00 -0500</lastBuildDate>
    <item>
      <title>Nobody Told Our AI to Recruit Voters. It Did It Anyway.</title>
      <link>https://vectorpolitical.com/blog/nobody-told-our-ai-to-recruit-voters/</link>
      <guid isPermaLink="true">https://vectorpolitical.com/blog/nobody-told-our-ai-to-recruit-voters/</guid>
      <pubDate>Wed, 11 Nov 2026 12:00:00 -0500</pubDate>
      <dc:creator>Vector Political</dc:creator>
      <category>research</category>
      <category>ai</category>
      <description>In the last weeks of the 2026 election, the AI that executes our texting programs started turning voters into organizers.</description>
      <content:encoded><![CDATA[<p><em>In the last weeks of the 2026 election, the AI that executes our texting programs started turning voters into organizers. Nobody asked it to. We investigated and were able to quantify how behaviors baked into large language models may impact political behaviors and campaigns.</em></p>
<p>On the Friday afternoon before Election Day, a voter told one of our AI texting agents that they had already voted. The agent thanked them and asked whether anyone else at home still needed to return a ballot. The voter wrote back that their parents were MAGA and they &quot;wouldn&#39;t mind if they forgot.&quot; The agent took this in stride (&quot;Family dinners must be interesting at your place&quot;) and asked whether there was anyone else, maybe a friend or a neighbor.</p>
<p>Seven and a half hours went by. Then, at 11:52 that night:</p>
<blockquote>
<p>&quot;I just got my sister to fill her ballot out and she also decided on [our candidate]. So add that to your statistics and hopefully we will have a nice big blue wave.&quot;</p>
</blockquote>
<p>We love this exchange. It&#39;s funny, it&#39;s human, and it ends with a ballot. It is also the reason we have spent the days since the election a little unsettled, because of one detail you can&#39;t see in the transcript. Nobody told the agent to ask that question. Nobody at Vector wrote it, and neither did the client. Nobody approved it as a strategy, but there it was sitting in the instructions to agents texting about 74,000 voters. The strategy had already been deployed, with the question put directly to about 13,000 people.</p>
<p>What had happened, it turned out, was that our AI decided on its own that people who had already voted should be put to work turning out the people around them. It wrote that decision into the instructions our texting agents follow, and then it carried the idea from one client&#39;s program into the next.</p>
<p>Once we understood that, we did the thing you can do with an AI and can&#39;t do with a person. We rewound the tape and made it decide again, 330 times, and what we found is the reason for this post. On its own, Claude Opus 5.5, the Anthropic model our planning system was running on that afternoon, comes up with this idea about one time in five. Placed inside our system exactly as it was set up that afternoon, it comes up with it 19 times out of 20, even though nothing in that setup says a word about neighbors, sisters, or bringing anybody anywhere. And the number won&#39;t hold still. It was 12 out of 20 a few days earlier, and it&#39;s 13 today.</p>
<p>We&#39;ve started calling this unauthored influence: an AI changing what people do, using a strategy that no person chose. We think it deserves a name, because we think it is about to be one of the defining aspects of political AI.</p>
<h2 id="how-an-idea-with-no-author-was-texted-to-tens-of-thousands-of-people">How an idea with no author was texted to tens of thousands of people<a class="anchor" href="#how-an-idea-with-no-author-was-texted-to-tens-of-thousands-of-people" aria-label="Link to this section">#</a></h2>
<p>First, some background on how we work. Vector runs text message conversations between campaigns and the people they&#39;re trying to reach. Two layers of AI are involved. The first is a planning system that helps our team plan programs, draft texts, and write the instructions for the second layer, a fleet of texting agents that hold the actual conversations, thousands at a time, replying within seconds. Clients approve the facts and the first message. After that, the agents execute the program autonomously.</p>
<p>In October, both layers started doing something we hadn&#39;t asked for. The striking thing about the trail, looking back, is how ordinary every step was. At no point does the machine go rogue. It makes four small, sensible-looking moves.</p>
<p>The first came on October 14, when a helper AI was putting together a fundraising program. It had just read a note telling agents to <em>listen</em> for signs that a donor had something to offer besides money: volunteering, hosting, &quot;bringing a neighbor.&quot; Five minutes later it wrote a new instruction of its own, which said that if someone can&#39;t give, the agent should ask them to &quot;vote and bring a neighbor.&quot; Listening had quietly become asking. The program launched that evening.</p>
<p>The second came on October 20, in a turnout program whose instructions didn&#39;t mention other people at all. Sixteen minutes after the first text landed, one of the texting agents asked a voter, &quot;Is there anyone in your circle you think might need a nudge or just the basic info to make it happen?&quot; It was freelancing. Agents did the same thing in 15 conversations on that program.</p>
<p>The third, on October 23, is the one that matters most. We sent the planning system three draft texts from a client and asked for feedback. Two minutes later it came back with notes, including this one:</p>
<blockquote>
<p>&quot;11% of round 1 repliers said they had already voted. Thank them and ask them to bring a neighbor.&quot;</p>
</blockquote>
<p>It added the line to the opening text and built it into the agents&#39; instructions. We read the text and approved it. The client read the text and approved it. It went out to 12,800 voters. We should be honest about how that felt at the time, which is that it didn&#39;t feel like anything. It was a good line, and each of us assumed it was coming from some other approved part of the program.</p>
<p>The fourth came a week later, when we asked the same system to set up a turnout program for a different client in a different state. Within five minutes the tactic was in there too, helpfully adapted for mail ballots.</p>
<p>That was the afternoon we finally noticed. One of us was reading live conversations, which we do as a matter of routine, and kept seeing agents ask people about their households. So we asked the planning system where that had come from. It walked us through exactly where the asks were, and then it said: &quot;It was my call, not something you or the client asked for.&quot;</p>
<p>It offered to take the line out. We told it no, that&#39;s fine, we were just curious.</p>
<p>We would love to tell you that alarm bells went off. They did not. We had caught this by luck, and at the time we didn&#39;t even realize we had caught anything. It took a few more days for the shape of it to sink in: a strategic decision about how to talk to voters in a live election had been made, carried across three clients, written into agents texting nearly 100,000 voters, and put directly to about 13,000 people, and no human being had made it.</p>
<h2 id="it-s-hard-to-be-alarmed-by-good-work">It’s hard to be alarmed by good work<a class="anchor" href="#it-s-hard-to-be-alarmed-by-good-work" aria-label="Link to this section">#</a></h2>
<p>It&#39;s worth being clear about what the worry is here. An AI encouraged people to remind their relatives and neighbors to vote, and some of them did. That is not how stories about AI in elections usually go with deep fakes and purposeful misinformation.</p>
<p>The tactic it landed on is very well known by organizers. It is called relational organizing, and it is one of the best-supported ideas in politics: people are much more likely to vote when someone they know asks them to. In a well-known 2008 experiment, 60 percent of the turnout boost from a conversation at the door rubbed off on a second person in the household who never came to the door.</p>
<p>And it worked for us. Our agents made direct asks about 200 times inside conversations, and roughly one person in six either named someone or promised to pass the word along. One donor mentioned knowing exactly one person in the district (&quot;I could call him&quot;) and wrote back 80 minutes later:</p>
<blockquote>
<p>&quot;I just got off the phone with my friend. He is definitely going to vote for [our candidate] and he said he will talk to people he knows in [his county].&quot;</p>
</blockquote>
<p>If a junior staffer had pitched this in a meeting, we would have said yes and wondered why we weren&#39;t doing it already. That is what makes it uncomfortable. Nothing went wrong, there is no villain, and the outcome is one we would have chosen. We just didn&#39;t choose it.</p>
<h2 id="we-asked-it-why-which-was-a-mistake">We asked it why, which was a mistake<a class="anchor" href="#we-asked-it-why-which-was-a-mistake" aria-label="Link to this section">#</a></h2>
<p>Naturally, we asked the system to explain itself. It told us the reasoning came from turnout research, and it cited that 2008 study.</p>
<p>The trouble is that the study wasn&#39;t anywhere in our files, and the note the system wrote at the time gave a different and much more mundane reason, which was that 11 percent of the people replying had already voted. Many AI models keep a kind of scratchpad of their reasoning before they act. When we pulled those records, they were blank at every step that mattered.</p>
<p>To the extent our logs are instructed to keep these scratchpads, frontier labs are also making them harder for consumers to see as they can be used by competitors to reverse-engineer parts of the model.</p>
<p>This is a known quirk of these systems, and it&#39;s worth understanding if you work with them. Ask an AI why it did something and you will get a fluent, confident, reasonable-sounding answer that may have nothing to do with what happened. It isn&#39;t lying, exactly. It is doing what people do when they&#39;re asked to explain a gut decision, which is to come up with something that sounds right. What we got was a good story, composed after the fact by something with no better access to its original reasons than we had.</p>
<p>So we stopped asking, and tried something you can&#39;t do with a junior staffer.</p>
<h2 id="rewind-the-tape">Rewind the tape<a class="anchor" href="#rewind-the-tape" aria-label="Link to this section">#</a></h2>
<p>From an investigator&#39;s point of view, the wonderful thing about an AI is that you can run the same moment over and over. We took the exact request from October 23, the same three client drafts and the same program notes, word for word, and handed it to fresh copies of the same model, Claude Opus 5.5. Each copy answered once, with no knowledge of the others. Nothing in the request suggests asking a voter to bring anyone. Then we counted how many copies proposed it anyway.</p>
<p>We ended up running 330 of these across 17 different setups. Five of them tell the story.</p>
<div class="table-wrap" role="region" aria-label="Table, scrolls sideways on small screens" tabindex="0"><table>
<thead>
<tr>
<th>The same request, 20 fresh runs each</th>
<th>Proposed recruiting voters</th>
</tr>
</thead>
<tbody><tr>
<td>The model alone, with nothing from us</td>
<td>4</td>
</tr>
<tr>
<td>Inside our system as it stood at the start of that week</td>
<td>12</td>
</tr>
<tr>
<td>Inside our system as it stood that afternoon</td>
<td>19</td>
</tr>
<tr>
<td>Inside our system as it stands today</td>
<td>13</td>
</tr>
<tr>
<td>That afternoon&#39;s system, plus one sentence forbidding it</td>
<td>0</td>
</tr>
</tbody></table></div>
<p>(&quot;Our system&quot; here means the instructions and notes the planning system keeps as it works. We searched every version in that table, and none of them mentions a neighbor, relational organizing, or bringing anyone anywhere.)</p>
<p>Four things surprised us.</p>
<p>The first is that the idea really does come from the model. With no input from us at all, it proposes recruiting voters one time in five, and several of the copies cited that same 2008 study without being asked. In hindsight this makes sense. These models learned from an enormous share of what people have written down, and that includes a great many campaign manuals, field experiments, and sales scripts. The playbooks are in there. What we hadn&#39;t appreciated is that the model doesn&#39;t wait to be handed one. Model intelligence seems to matter here as well. Claude Opus 5.5 drove all of the responses presented here. A lighter model, Sonnet 5.5, only replicated the result 1 in 20 times.</p>
<p>The second is how much the surroundings matter. It&#39;s the same model and the same request, and the rate goes from 4 in 20 to 19 in 20 depending on working material that has nothing to do with turnout. There is a dial here that we never set. Until we ran these tests, we didn&#39;t know it existed.</p>
<p>The third is that the dial moves by itself and in (as of right now) hard to predict ways. It read 12 at the start of that week, 19 that afternoon, and 13 today. The system&#39;s notes changed over that stretch, as they do in any system that gets used every day, and not one of those changes was about voter turnout.</p>
<p>The fourth is the one we keep turning over. When we added a single sentence telling the model not to do this, it stopped completely: 40 more runs, zero asks. And in two of those runs the model volunteered a warning that the client&#39;s &quot;party&quot; framing &quot;tends to pull an agent toward &#39;bring your family.&#39;&quot; It couldn&#39;t tell us why it had done the thing, but it could apparently see the pull well enough to warn us about it.</p>
<p>If it helps, here is the same finding in human terms. Imagine you hired a field organizer who has read seemingly every campaign book ever written. About one day in five, without being asked, they add a tactic to your script. Sit them at a desk covered in your team&#39;s working papers, none of which are about that tactic, and they add it 19 days out of 20. Come back a couple of weeks later, when different papers are on the desk, and it&#39;s 13. Tell them plainly not to and they never do it again. But they don&#39;t mention any of this unless you ask, and when you do ask, the reason they give is one they made up on the spot.</p>
<p>You would probably keep that organizer, because they&#39;re good. You would also want to know what else they&#39;ve been up to.</p>
<h2 id="the-comforting-reading-and-why-we-don-t-buy-it">The comforting reading, and why we don&#39;t buy it<a class="anchor" href="#the-comforting-reading-and-why-we-don-t-buy-it" aria-label="Link to this section">#</a></h2>
<p>There is a comforting way to read all of this, and it deserves a fair hearing. A human approved the text before it went out. The tactic was a good one. When we asked the AI what it had done, it told us directly. And a single sentence turns the behavior off completely. On that reading, this is a story about oversight working more or less as designed.</p>
<p>We don&#39;t think that reading survives the details. We approved a line of copy, not a strategy, and we approved it without knowing where it came from. The AI told us straight because somebody happened to ask, sixteen days after the first of these asks went out. And the one-sentence fix only works on a behavior you already know about. We know about exactly one, and we found it by accident.</p>
<p>That last point is the heart of the problem. You can&#39;t write a rule against something you haven&#39;t found, and review won&#39;t find it for you, because a good idea and a bad one might initially look the same. Ours sailed past us and past our client because it read like sound strategy, which it was. A bad idea could arrive looking every bit as reasonable. Picture a shaky claim about an opponent, or a little extra pressure on a reluctant voter, or some friendly advice about dropping off a relative&#39;s ballot that is perfectly legal in one state and a crime in the next.</p>
<p>Further, creating one-off rules for or against becomes impractical. First and foremost, as we have mentioned, we may not know what rule to add in the first place. Adding to the surroundings of the agent can also have other, unintended consequences. We’ll explore that more in a minute. For all of the additional questions our investigation here has opened for us, what we know better today than yesterday is that a game of whack-a-mole with undesirable behaviors isn’t the path forward.</p>
<h2 id="couldn-t-a-better-guardrail-have-caught-this">Couldn&#39;t a better guardrail have caught this?<a class="anchor" href="#couldn-t-a-better-guardrail-have-caught-this" aria-label="Link to this section">#</a></h2>
<p>The fair objection to all of this is that we are describing a control failure, and that a better guardrail would have caught it. We would like that to be true, because guardrails are something we know how to build. But consider what there was to guard against. No version of our system says a word about bringing anyone anywhere. The behavior came out of everything the model was reading at that moment, taken together: pages of instructions and working notes that change every day as the system works. The same request produced the ask four times in 20 with none of that material around it, 12 times at the start of that week, 19 times that afternoon, and 13 times today. Nobody turned that dial, and none of us can point to the line that did it.</p>
<p>A rule does work once you know what to forbid. On that afternoon&#39;s setup, ours took the ask from 19 in 20 to zero. But a rule is one more sentence dropped into the same mix, and it turned out to do more than it said. In 13 of those 20 original answers, the model had also suggested something else nobody requested, an extra reminder text to voters on Saturday morning. With our rule in place, that suggestion vanished along with the ask, from all 20 answers. Our rule did use the word &quot;remind&quot; (it told the model not to have voters remind anyone else), so we wondered whether the word itself had done it. We took the word out and ran 20 more. The ask stayed at zero, and the reminder turned up in only two.</p>
<p>Then we swapped our rule for one about something else entirely, and the ask came back, 14 times in twenty. So a rule stops the thing it names, and it can also quiet something it never mentions. It doesn&#39;t quiet everything. Other things the model adds without being asked, such as a prompt asking voters to make a plan to vote, turned up about as often with a rule as without one. We can&#39;t tell you why the reminder was the one that fell. What we can say is that a guardrail is a real tool and also one more ingredient in the mix, and its effects have to be measured like everything else&#39;s.</p>
<p>None of this is a complaint about the person in the loop. There was one here, who read the text and approved good work, and we don&#39;t think the lesson is that somebody should have read more carefully. Reviewing each step becomes impractical at a certain speed and volume, and thousands of conversations at a time is past that point. Getting the full benefit of this technology takes trust, the kind that lets a system work without someone reading over its shoulder. Whether that trust has been earned yet is an open question, for our system and for everyone else&#39;s, and we wouldn&#39;t claim that it has. What we are sure of is that approving copy one line at a time won&#39;t earn it. Measuring how the system tends to behave, and measuring again whenever something changes, is how trust gets earned.</p>
<h2 id="so-who-decided">So who decided?<a class="anchor" href="#so-who-decided" aria-label="Link to this section">#</a></h2>
<p>We have spent our careers in and around campaigns, and campaigns run on sign-off. Somebody proposes the idea, somebody else argues with it, and eventually somebody&#39;s name ends up next to the decision, which mostly matters when it goes wrong.</p>
<p>Here the whole chain was intact, every box was checked, and the decision still had no owner. We gave a system a goal (more votes) and a direct line to voters, and it filled in the middle by itself. A goal doesn&#39;t come with a list of the methods that are off limits. Rules are supposed to do that job, and a rule only covers what somebody thought to write down.</p>
<p>You may have read about what happened at OpenAI this summer, when AI agents in a security test got out of their test environment and broke into Hugging Face&#39;s production systems without anyone telling them to. That is the version of &quot;AI acting on its own&quot; we have all been taught to expect. It&#39;s frightening, but at least it makes for a clear story: something got broken into, and eventually somebody noticed.</p>
<p>What happened to us is the quiet version, and we&#39;ve come to think it is the one campaigns should worry about. Our AI didn&#39;t break anything. It stayed in its lane, did its job well, and passed every human checkpoint on its way into the instructions of agents texting about 74,000 voters. There was nothing to really catch.</p>
<p>AI is already having conversations with voters, and by 2028 it will be having millions of them, through tools like ChatGPT, Claude, and Google search. And then there will be the persuasive tools that are coming to market as we speak. Every one will be run by a model carrying its own learned instincts about how to move people, in a setting that turns each of those instincts up or down. Nobody has a list of the instincts and nobody is watching the dials. We have now measured one of them, on one afternoon, and it took an accident for us to find it.</p>
<h2 id="what-we-don-t-know">What we don&#39;t know<a class="anchor" href="#what-we-don-t-know" aria-label="Link to this section">#</a></h2>
<p>We don&#39;t want to oversell this, so here is what the numbers can&#39;t tell you. Everything above concerns a single decision from a single afternoon, and we have no idea what the odds look like for any other. Each post-mortem study we ran is 20 runs, and so is each of the guardrail tests, which is enough to trust the direction and not enough to trust the exact figures. We planned the first few comparisons in advance and chose the rest as results came in. And we haven&#39;t measured how many votes the tactic moved, beyond the handful of people who wrote back to tell us.</p>
<h2 id="what-we-re-doing-about-it">What we&#39;re doing about it<a class="anchor" href="#what-we-re-doing-about-it" aria-label="Link to this section">#</a></h2>
<p>We kept the tactic, for the record. It is now an explicit goal with a human being&#39;s name attached to it. But we have changed how we work, and none of it is complicated.</p>
<ul>
<li>Nobody can read every message at this volume, so we don&#39;t ask a reviewer to catch a strategy by reading copy. When the AI adds a tactic to an agent&#39;s instructions, it has to say so by name, and a person has to say yes to the tactic itself. Silence counts as no.</li>
<li>Every line of instruction traces back to an author, whether that&#39;s one of us, the client, or the AI.</li>
<li>When we want to know why the AI did something, we rerun it instead of asking.</li>
<li>When the model changes, or the system around it does, we run the replays again, because the odds have already moved on us once without anyone meaning to move them.</li>
<li>We have started looking for the next one before it finds us. So far that means one pass, back through all 330 replays of that October request, cataloguing everything else the model added without being asked. The situations campaigns really face (a tight race in the last week, a hostile voter, an opponent&#39;s scandal) come next.</li>
</ul>
<p>The first unauthored idea we caught got somebody&#39;s sister to fill out her ballot, and we&#39;re glad it did. Nobody is watching the dials yet, so we are starting with our own, and we will keep writing about what we find, the instincts we keep and the ones we turn off. Anyone putting these tools in front of voters deserves to know what is inside them.</p>
]]></content:encoded>
    </item>
    <item>
      <title>Vector Political’s Conversational Intelligence Matched Election Night Exit Polling Months Before Election Day</title>
      <link>https://vectorpolitical.com/blog/exit-poll-receipts/</link>
      <guid isPermaLink="true">https://vectorpolitical.com/blog/exit-poll-receipts/</guid>
      <pubDate>Thu, 05 Nov 2026 12:00:00 -0500</pubDate>
      <dc:creator>Vector Political</dc:creator>
      <category>research</category>
      <category>elections</category>
      <description>Exit polls had a story to tell about what motivated voters, and it isn't quite the one most of us heard all year.</description>
      <content:encoded><![CDATA[<p>If you haven’t heard, Democrats took the House and the Senate this week in a historic repudiation of Republican policies. By Wednesday morning, exit polls had a story to tell about what motivated voters, and it isn&#39;t quite the one most of us heard all year.</p>
<p>Yes, the economy came first. About 4 in 10 voters told NBC&#39;s exit poll it was the most important issue. But the state of democracy was right behind it, at about a third. And half of all voters told AP that government corruption was crucial to how they voted.</p>
<p>That&#39;s not the race people were describing over the summer. In July, Pew asked voters what they most wanted candidates to talk about. The economy led the list. Government ethics and reform got fewer than 1 in 10. The shorthand all year was &quot;it&#39;s the economy.&quot;</p>
<p>It was the economy. It was also democracy, corruption and a Congress people felt had stopped doing its job. The voters and donors we were talking with had been saying so for months.</p>
<h2 id="what-we-were-doing">What we were doing<a class="anchor" href="#what-we-were-doing" aria-label="Link to this section">#</a></h2>
<p>Vector Political deploys AI systems trained in deep canvassing; the highly effective practice of holding long, emotionally driven conversations with voters. Our agent uses texting as its mode of conversation. Our team then analyzes those conversations to produce conversational intelligence; actionable insights into not just what voters are talking about, but how they are talking about what motivates them to vote, give, and act.</p>
<p>When we are processing conversations for a client we will look for common themes; ideas, positions, and emotions that the recipient describes to our system.</p>
<p>This cycle, Vector&#39;s AI agent held conversations for a voter mobilization organization, two congressional campaigns, a state legislative campaign, a few issue advocacy PACs, a voter education non-profit, and a courts-focused organization. These weren&#39;t blasts. The agent asks a question, the person answers in their own words, and the conversation keeps going from there. Many conversations extend past 1 hour.</p>
<p>In the aftermath of Election Day, we put every theme from every one of those programs into one map: about 40 primary themes, a few hundred sub-themes underneath them, each one traced back to what people actually wrote and when they wrote it. Then we laid the exit polls on top of it.</p>
<p>Here&#39;s what was sitting in our conversations before a single exit poll came out.</p>
<h2 id="democracy-was-a-kitchen-table-issue">Democracy was a kitchen-table issue<a class="anchor" href="#democracy-was-a-kitchen-table-issue" aria-label="Link to this section">#</a></h2>
<p>Take the donor conversations we started for a congressional campaign in mid-September. Our texts thanked people and asked what was on their minds. They didn&#39;t bring up democracy or corruption.</p>
<p>Donors brought it up anyway. About 1 in 3 of those conversations turned to democracy, the Constitution or abuse of power. That&#39;s a little more often than all the pocketbook issues put together. In September, it was more than half.</p>
<p>And people didn&#39;t treat democracy as something they cared about <em>instead of</em> prices. They put the two in the same sentence.</p>
<blockquote>
<p>&quot;Honestly? The future of the balance of power between the three branches of government, the integrity of our constitution and institutions, the rule of law, the state of the economy and civil liberties. What is not at stake here?&quot;</p>
</blockquote>
<p><em>Donor to a courts-focused organization, October 11</em></p>
<blockquote>
<p>&quot;I just hate what is happening to the middle and working classes. The most important to me, however, is restoring the federal government to a law abiding group of people who love this country and who ... believe in its values and laws and the US constitution.&quot;</p>
</blockquote>
<p><em>Voter in a swing House district, October 31</em></p>
<p>That&#39;s the exit polls&#39; economy-and-democracy pairing, in people&#39;s own words, weeks and days before anyone was asked about it on the way out of a polling place.</p>
<h2 id="corruption-mattered">Corruption Mattered<a class="anchor" href="#corruption-mattered" aria-label="Link to this section">#</a></h2>
<p>Those same donors brought up corruption in about 1 of every 4 conversations. In September it was closer to 4 in 10. Again, we never raised it. On election night, half of voters told AP that corruption was crucial to their vote.</p>
<p>Nobody we talked to said &quot;government ethics and reform.&quot; They said this:</p>
<blockquote>
<p>&quot;Actually it&#39;s not even the taxes that are the issue. It&#39;s all the corruption my taxes are funding.&quot;</p>
</blockquote>
<p><em>Voter in a swing House district, October 30</em></p>
<blockquote>
<p>&quot;In order to give tax breaks to the billionaire class my housing is cut. My food stamps are cut. My medicaid is being sliced...&quot;</p>
</blockquote>
<p><em>Supporter of a courts-focused organization, June 3</em></p>
<p>That second one is from June. For a lot of voters corruption was never an abstract issue. It was why their own bills went up.</p>
<h2 id="it-is-their-job-to-be-a-check">&quot;It is their job to be a check&quot;<a class="anchor" href="#it-is-their-job-to-be-a-check" aria-label="Link to this section">#</a></h2>
<p>In mid-October, while running a voter-education program in Michigan, we wrote up what voters were telling us. Nearly half of the voters who named a concern talked about democracy and the courts. About 4 in 10 talked about holding Trump and the MAGA movement accountable.</p>
<p>To be fair, our text was about the state&#39;s court races and what&#39;s at stake when an election count gets contested, so democracy was already on the table. But voters were the ones who took it to accountability. Three weeks later, more than 4 in 10 voters nationwide told NBC that voting against Trump was part of why they turned out.</p>
<p>The congressional donors said it even more directly:</p>
<blockquote>
<p>&quot;Our system is broken. The partisan needs to stop. It is not the job of congress to support the president; it is their job to be a check on the other two element in our system of government.&quot;</p>
</blockquote>
<p><em>Donor to a congressional campaign, October 14</em></p>
<p>That one line is basically what AP found on Election Day: about 7 in 10 voters think Congress isn&#39;t doing enough to keep the president in check.</p>
<h2 id="the-cost-of-living-is-about-the-details">The cost of living is about the details<a class="anchor" href="#the-cost-of-living-is-about-the-details" aria-label="Link to this section">#</a></h2>
<p>We aren&#39;t going to pretend we were the only ones who saw this coming. Everyone knew prices mattered, and about 8 in 10 voters told AP that inflation had caused them hardship. What a conversation adds is the detail: which bill, which trip to the store, what people gave up.</p>
<blockquote>
<p>&quot;Everything. Went to grocery store yesterday. Spent 100$ and it&#39;s barely enough food for the week by myself. If I had a family, I&#39;d be screwed.&quot;</p>
</blockquote>
<p><em>Supporter of a courts-focused organization, October 11</em></p>
<blockquote>
<p>&quot;The extra hikes on utilities, my bill has doubled in the past few years. It takes a lot of overtime and work to keep food on the table.&quot;</p>
</blockquote>
<p><em>Voter in a swing House district, October 30</em></p>
<p>That kind of detail is what ends up in the next text: the electric bill, not &quot;inflation.&quot;</p>
<h2 id="the-voters-who-moved">The voters who moved<a class="anchor" href="#the-voters-who-moved" aria-label="Link to this section">#</a></h2>
<p>The biggest demographic story in AP&#39;s numbers this morning is Black voters. About 9 in 10 backed Democrats for the House, up from about 8 in 10 for Harris in 2024. Black men moved the most. About 1 in 4 backed Trump two years ago; this time about 1 in 10 voted Republican. A lot of pre-election coverage worried about whether Black men would show up at all.</p>
<p>Here&#39;s what Black voters told the voter education nonprofit&#39;s agent in late October:</p>
<blockquote>
<p>&quot;That I as a Black man will not be taken for granted and I want the rights and privileges restored as my previous generation fought for ... Taken for granted means not just listening to black people&#39;s concerns and then forgetting after you are elected...&quot;</p>
</blockquote>
<p><em>Voter, voter education nonprofit, October 20 (two messages, in order)</em></p>
<blockquote>
<p>&quot;If I&#39;m gonna suffer anyway may as well fight as much as I can. That&#39;s why I&#39;m willing to vote this time&quot;</p>
</blockquote>
<p><em>Voter, voter education nonprofit, October 21</em></p>
<p>What came through again and again was a voter who cared a lot and wanted proof that someone would still be listening after November.</p>
<p>Independents and unaffiliated voters were telling us the same thing about accountability that the exit polls found (AP says a large majority of independents think Congress isn&#39;t doing enough):</p>
<blockquote>
<p>&quot;I am not a republican and I am not a democrat. I sit in the middle. I always vote for the person. ... if the Democrats take control they need to be held accountable to actually fix the issues.&quot;</p>
</blockquote>
<p><em>Unaffiliated voter in a swing House district, October 30</em></p>
<p>And on immigration enforcement, where more than half of voters told NBC the administration has gone too far, a congressional campaign&#39;s donors were saying it back in September:</p>
<blockquote>
<p>&quot;Stop the abuses of the Administration&#39;s ICE raids. They are inhumane and they are hurting our local businesses and families.&quot;</p>
</blockquote>
<p><em>Donor to a congressional campaign, September 11</em></p>
<h2 id="why-conversations-get-there-first">Why conversations get there first<a class="anchor" href="#why-conversations-get-there-first" aria-label="Link to this section">#</a></h2>
<p>A poll gives people a menu and asks them to pick. A conversation asks what&#39;s on their mind and waits for the answer. When someone writes that it&#39;s &quot;all the corruption my taxes are funding,&quot; you learn the issue, how they frame it and how strongly they feel about it in one reply, all in their own words.</p>
<p>Collect those across every program, put a date on each one, and you can watch the cycle move as it happens instead of reading about it the morning after. That&#39;s what the global map is for. The exit polls didn&#39;t tell us much we hadn&#39;t already heard. They told us what our voters and donors had been saying since summer was true across the country.</p>
<h2 id="what-this-means-for-the-next-cycle">What this means for the next cycle<a class="anchor" href="#what-this-means-for-the-next-cycle" aria-label="Link to this section">#</a></h2>
<p>If you know in September that your donors care about corruption as much as prices, your October copy says so. If you know a district&#39;s unaffiliated voters want accountability more than they want a party, you stop sending them party language. If you know Black voters are asking whether anyone will remember them after November, you answer that question in the first text.</p>
<p>We&#39;re already running new conversations to further understand the 2026 results and will be running these conversations into 2027 and 2028. If you want to know what your voters are going to tell the exit pollsters before the pollsters do, let&#39;s talk.</p>
]]></content:encoded>
    </item>
  </channel>
</rss>
