What You Should Expect of Your Campaign AI

It's 9 PM on a Tuesday. A state legislator just introduced a bill that would directly impact thousands of families in your district. Your opponent is silent. Your supporters are angry. This is the fundraising email that writes itself.

Except nobody on your team saw it yet. Your comms director will find it on Twitter tomorrow morning. You'll have a draft by Thursday. By Friday, your opponent has already sent their version and raised $40,000 off it.

That three-day gap used to be normal. In 2026, it's a choice.

Most campaigns that say they're "using AI" mean someone on staff has a ChatGPT tab open. That's not a campaign AI. That's a search engine with better grammar.

Here's the bar for what your AI should actually be doing.

Your AI should catch the story before your team does.

A bill drops at 9 PM. Your AI should have flagged it by 10. Not because someone set up a Google Alert for that specific bill number. Because the AI is scanning legislative trackers, district news outlets, state capitol coverage, and community sources in the languages your district actually speaks. Every few hours. Scoring every development against your candidate's story and your campaign's messaging.

It should surface it with context: "This connects to the education funding issue you've been hitting. Your last two emails were policy-heavy, so this one should lead with the personal angle. Here's a draft."

By the time your comms director checks Twitter in the morning, your draft is already waiting for approval on the candidate's phone.

Your AI should write like the candidate talks. Not like a committee drafted it.

Not "Dear Supporter, in these challenging times, it's more important than ever that we stand together for our shared values." Every donor in America can smell that from the subject line.

Your AI should know the candidate's real voice. How they tell the story about why they got into this. The way they talk about their community versus the way a policy staffer would write about it. When a bill drops about education funding, the AI shouldn't produce a summary. It should write the email the candidate would write at midnight after reading the bill and getting angry about it.

A donor reads it and thinks "that sounds just like them." That's the standard. If it reads like it was generated, it's not ready.

Your AI should know that blasting 30,000 people at once is malpractice.

Here's what happens every cycle. Campaign writes a decent email. Sends it to the entire list at 9 AM. 85% never open it. Gmail notices. Next email goes to spam for half the list. Campaign wonders why open rates dropped. Sends another blast anyway. Repeat until the list is dead by October.

Your AI should send to your most engaged supporters first. A few thousand. Watch how they respond. If they're opening, clicking, donating: expand to the next tier. If they're not: stop and figure out why before you burn the other 25,000.

This single decision — tiered sending versus blasting — is the difference between campaigns struggling at 15% open rates and campaigns performing at levels most political operatives wouldn't believe. It's not better copy. It's smarter delivery.

Your AI should make sure no donor slips through the cracks.

Candidate's at a Lincoln Day dinner. Someone walks up, hands over a business card, says "I maxed out to the last guy who ran. I want to help." That card goes in a jacket pocket. The jacket goes in the car. The card sits on a kitchen counter for a week. Eventually it ends up in a drawer with thirty other cards from thirty other events.

That person wanted to write a check. Nobody followed up.

Your AI should turn that business card into a donor profile before the candidate leaves the parking lot. Photo of the card. Enriched with public donation history, giving patterns, connections to other supporters. In the CRM. Tagged. A personalized follow-up queued for 48 hours later referencing the event and the conversation.

Every campaign manager in America says they need better donor follow-up. They don't. They need a system that doesn't depend on someone remembering.

Your AI should know what your opponent did yesterday.

New committee filing hits the public record at 11 PM? Your AI should surface it before breakfast. Opponent takes a position that contradicts something they said six months ago? Logged, with receipts. An independent expenditure group files in your race? You should know within hours, not when a reporter calls asking for comment.

This isn't opposition research. That's still humans doing careful, sourced work. This is automated awareness. The difference between being surprised and being prepared.

Everything should talk to everything.

The intelligence layer should tell the content engine what to write. The email system should know what the CRM knows. When a donor clicks a link about veterans' issues, the next email they get should reflect that. When the news scanner flags a story, the draft should already account for which supporters care most about that topic.

If your campaign runs on six disconnected tools connected by a staffer copying and pasting between browser tabs, you don't have AI operations. You have a person doing data entry with a nicer font.

Here's the honest part.

Most campaigns in 2026 won't hit this bar. Most will use AI the way campaigns used social media in 2012: tentatively, partially, without understanding what it actually does. A lot of races will be won and lost the old-fashioned way.

But some campaigns will get this right. Not a chatbot. Not a fancy typewriter. A system that watches, writes, decides, remembers, and coordinates while the candidate does what only the candidate can do.

The gap between those two kinds of campaigns is going to be the story of this cycle.

Filing deadline is March 6.

The infrastructure you launch with matters more than the infrastructure you wish you had six months from now.

Eric Linder is the founder of AutomatedTeams, a political AI operations consultancy. Former California State Assemblyman (2012-2016). He builds and operates AI agent systems for political campaigns, PACs, and government affairs organizations.