Trends

Iowa Is Too Late: Why 2028 Presidential Campaigns Need to Start on AI Now

Nov 16, 2026
7 min read
Iowa Is Too Late: Why 2028 Presidential Campaigns Need to Start on AI Now

The 2028 primaries are more than a year away but the version of your candidate that AI presents to primary voters is being formed now.

The 2028 presidential primaries are more than a year away but the version of your candidate that AI presents to voters is being formed right now. Beyond a candidate website, most presidential campaigns won’t think about their digital strategy until the early states are already in sight but by then, much of what AI knows about their candidate may already be set.

Given that President Trump can’t run for a third term and the field for Democrats will be split between moderate veterans and progressive upstarts, both parties will likely face prolonged, contested nominations and every prospective candidate will need to rise above political noise to introduce themselves, their record and their positions to prospective voters.

Hundreds of millions of dollars will be spent over the course of the nominating campaigns but the channel that will likely give prospective nominees the best ROI is increasingly the first place voters will meet them: AI.

An estimated 123.5 million Americans voted in the 2026 midterm elections and according to our internal polling data, Rankbee believes AI influenced between 2.4 and 2.8 million votes. Given the ubiquitous nature of AI and the speed of adoption and acceptance, we believe those figures will be substantially higher in 2028 and the sooner candidates and their teams prepare for its influence, the better they will be positioned for the brusing fight ahead.

How The 2026 Midterm Election is Already Shaping 2028

The 2026 midterms delivered a major shift at the state and federal level with Democrats winning key majorities in the House and Senate as exit polls showed voters held strong concerns over the economy and the direction of the country and felt Republicans weren't doing enough to address key issues they were empowered to fix following the end of the Biden administration.

Traditional deep red states like Mississippi, South Carolina and Texas elected Democratic Senators and key Trump allies like Greg Abbott and Joe Lombardo lost their gubernatorial races.

While the night was clearly one of celebration for Democrats across the country, the affordability struggle that defined this election cycle is unlikely to see meaningful change. Over the next two years rising debt levels are expected to keep interest rates high and diesel prices will likely remain elevated until a permanent solution is in place for the Middle East. White-collar hiring is expected to remain muted as companies continue to assess the impact of AI on headcounts and these factors combined with President Trump’s veto power and likely unwillingness to work with Democrats on legislation that gives them a signature 'win' before 2028 means affordability is unlikely to improve.

Thus the economy will once again dominate the next presidential election cycle and gridlock over the next two years will open the door for Republicans to win back disenfranchised voters who will be struggling with high costs.

Presidential candidates who plan ahead and establish themselves early through a strong digital identity will reap the benefits of an even stronger AI presence which is critical for early state voters.

How AI 'Forms' Political Views

Before we go further, lets start with what LLMs are and aren’t doing when a voter asks about a candidate's political position or compares candidates against one another.

AI is designed to be balanced, not to pick sides. Every LLM from ChatGPT, to Claude to Gemini and Grok are governed by instructions users almost never see called system prompts. These prompts tell them to be friendly, helpful, thorough and gate what information they can and cannot provide to users.

For political questions, those instructions generally tell models to fairly lay out the candidates and their positions and leave political conclusions and decisions to the user. Models aren’t perfectly neutral, and researchers have documented political leanings in some of them, but no mainstream assistant is purpose-built to intentionally steer voters toward a party or political belief both as a matter of principle and regulatory capture.

AI will not completely favor one candidate over another, so campaigns and their communications teams need to do is make sure that when it describes a candidate it has accurate, persuasive and complete material to work from.

The Two Key Sources of Political Information

  1. What the model learns from training data.

Large language models learn from an enormous collection of text including books, websites, articles and other written material, gathered up to a specific date called the “knowledge cutoff.”

As the model is fed billions of new content pieces it picks up the patterns of language and the facts and ideas that show up again and again. These patterns become a knowledge base for candidates and codify their underlying politics. However the process of training a model takes time and signifiant compute demands so it is done in large batches to maximize cost efficiency and effectiveness. Put bluntly, while models are constantly learning, they are not constantly updating

This means that information, or lack thereof, regarding a candidate, their political history, positions, voting record etc ... is by design out of date with anything that has recently happened. Thus a candidate declaring their intention to run for president today won't show up as a fact in training data until months later.

  1. What the Model Sees in Realtime on the Web

Knowing that training data is out of date by design, LLMs have built in systems that identify when a user's prompt is unlikely to be accurate when answered only with training data and in those cases it will search the web for information to supplement a response. This system creates a best case scenario where training data has a strong foundation and web information is used to build the house on top of it.

Web search pays off right away in the sense that a press release, interview, site update or social media statement can show up quickly in AI answers but training data pays off later and for longer.

Understanding Your Candidate's Views vs AI

The goal for any campaign team today needs to be understanding what information AI and by extension the web has on your candidate in relation to key issues. Communications teams need to pay attention to the accuracy of that information and more importantly where AI has gaps.

"Gaps can create major obstacles for campaigns," says Aris Vrakas, founder of Rankbee "If an AI model lacks clear information about your candidate's position on housing, tariffs or Middle East policy for instance, it may use fringe sources to build an inaccurate conclusion."

Gaps are perfectly normal. A state representative running for a national position is unlikely to have on-the-record views about foreign policy or trade agreements and a political outsider deciding to run for office will have little to no information on political stances other than what may be inferred from their non-political life.

Presidential candidates will likely have existing information and voting records that provide AI with a foundational layer covering most of the key voter issues but it's the grey areas around state specific concerns or motivators for specific voting blocs that may need more input.

The AI Presidential Calendar for 2028

Campaigns generally treat the months before the Iowa Caucus and early primary states as the place where positioning, differentiation and messaging begins. AI runs on a different calendar.

Here's Rankbee's practical timeline for how campaigns need to think about AI and its role in the nominating process:

  • Now to Mid-2027: Build the foundation for your candidate.
    • Map the questions voters will ask AI: Start with the major issues first, then consider background questions like “Who is this person?” and “What have they done?” If your campaign needs help mapping these out, our political marketing team can research and contextualize everything you need to consider.
    • Publish clear, specific positions on your own site first: In plain language and being specific about your candidate's view of the current state and what policies they would change and what they would continue. Vague wording on key topics will produce vague AI answers.
    • Tell the whole story: Background, record, why they’re running and what they’d change. That’s the narrative AI will reflect back to voters.
  • Mid-2027 to Late 2027: Know your presence and fill the gaps.
    • Understand your baseline: Once everything has been published, start to monitor what AI is saying about your candidate, both at a high level and contextualized to specific voting blocs by region.
    • Solve inaccurate information: If answers about your candidate do not accurately reflect their current position, determine if the issue is a lack of available sources or
    • Build differentiation: As the candidate field firms up, anticipate AI having to give voters a compare and contrast between candidates and creating content that clearly advocates your candidate as having a superior approach will create more compelling AI answers.
  • Primary season and beyond: Manage the live record.
    • Keep positions current: As events unfold, it is incredibly important to keep your campaigns owned assets up to date. Foreign policy decisions and new economic challenges will evolve narratives and when voters want information and it's critical they have the right answer
    • Respond to challenges: Rival campaigns both among your party and opposition will seek to attack voting records, changes in political views and candidate backstories. Addressing attacks with easy to find information for humans and AI will help to alleviate the impacts of hostile narratives.
    • Evolve as necessary: Should your candidate fail to make an impression as a presidential candidate, it's important to manage their online image for their next aspiration (VP consideration, cabinet secretary roles ...) Depending on circumstances and retention, you'll want to make sure their presence reflects aspirations as they evolve.

Given the likely size and variety of candidates for both Democrats and Republicans, every percentage point counts and an edge in AI can help campaigns cut through the noise and really reach voters.

The bottom line

The 2028 primaries are more than a year away but the version of your candidate that AI presents to primary voters is being formed now.

Whether you're already working on messaging or in the early stages of exploration, teams need to ensure AI can describe your candidate accurately. If your positions aren’t clearly on the record, AI will describe them poorly or not at all. Campaigns that build that record early will start the primaries with an advantage that late starters can’t buy back.

RankBee can help. Our team of political consultants can audit what AI says about your candidate today, build a plan for what to publish and where, and monitor how that picture changes through 2028. Get in touch to start planning.

Photo by Anival Torres on Unsplash

Never Miss an Article

Get weekly tips, case studies, and AI insights delivered to your inbox.