Dispatches from the intersection of political campaigns and AI operations.
The honest retrospective. What worked, what overpromised, what surprised, and what the campaigns that are already thinking about 2028 should be building now.
October is the biggest fundraising window of the general election cycle. The system that runs it needs to be at full capacity before October 1, not during it. Here's what sprint-ready looks like.
October is 30 days away. The AI operation you have right now is the one you're running with. Five things to check before the push starts.
By September, every fundraising ask feels like a September ask: urgent, transactional, and one of a dozen the donor is fielding. August is the last month where a conversation still feels like a conversation. Use it.
Every campaign AI operation that runs correctly in June has failure modes under sprint conditions. Three specific things break consistently, and none of them require a technology fix.
The campaigns that deployed AI in Q1 and Q2 have six months of compounding behind them. In August, that gap stops being theoretical. Here's the honest picture of what starting in August means, and what's still possible.
The final 100 days aren't a time to configure. They're a time to execute. The campaigns that win this window have already built the operation. Here's what executing at full capacity looks like.
Seven items. If any of these aren't checked before August 1, that's your configuration gap going into the sprint. Here's what a fully operational AI campaign looks like at the starting line.
Campaign phone banking automation gets a lot of hype. Here's the honest breakdown of what AI voice actually handles in a real campaign operation, and where it still needs a human.
The question is coming. 'Did your candidate actually write this?' Here's the honest answer, why it's not a problem, and how to handle it when it comes up in person.
August through October is the final fundraising push before November. Campaigns that haven't built their AI operations by July will be configuring under pressure in September. Here's what 'fully operational' looks like.
If you deployed AI in Q1 or Q2, six months of drift has probably accumulated. Voice models that haven't been updated. Approval workflows running slower than they should. Here are five questions to run against your operation before the sprint starts.
Summer feels slow for political fundraising. It isn't. It's the cultivation window that determines who shows up in October. The campaigns that understand this use June and July differently than everyone else.
The campaigns that deployed AI in Q1 and Q2 have something you don't yet: five months of compounding. Voice models trained on real conversations. Approval workflows that run without friction. Here's what that gap looks like and whether it's closable.
The campaigns that deploy AI badly have one thing in common: they had the wrong mental model before they started. Here are the five mistakes I see most often.
Your opponent isn't watching more news than you. Their system is. Here's what continuous AI news monitoring looks like for a campaign that doesn't have a full research staff.
The primary season created a natural experiment: some campaigns deployed AI operations correctly, some deployed incorrectly, and some didn't deploy at all. Here's what actually showed up in how campaigns ran.
DeepSeek and other Chinese AI models are genuinely capable and cost nothing to run. Campaign managers are starting to ask. Here's the honest answer.
Campaign managers ask us all the time: what does this actually look like in practice? Here's the complete workflow, from call session to approved batch, with nothing glossed over.
A business card is just a name and a number. Or it's a complete picture of who this person is, who they know, and how to make the ask that lands. The difference is 30 seconds.
Every vendor is selling AI now. The pitch sounds identical regardless of what the software actually does. Here's what's actually working in 2026, what isn't, and how to tell the difference.
Campaign fundraising runs on two things: relationships and volume. AI doesn't replace the relationship. It handles the volume so the relationship can happen.
Campaign AI gets covered as a technology story. The actual question is operational: which specific problems does it solve, and which ones does it still need a human for? Here's the map.
Most campaigns using AI are only scratching the surface. Here's what your campaign AI should actually be doing in 2026, from news scanning to donor follow-up to opponent tracking.
Campaigns lose donors to dropped follow-ups every week. Here's the anatomy of how commitments die, and how campaign follow-up automation catches every one before the 72-hour window closes.
A pet grooming business in a small California town went from 3 inbound calls a day to 30. No ads. No paid traffic. Just SEO done properly. Here's what that means for your campaign.
That donor database you're so proud of? It's a glorified spreadsheet with better branding. Here's what's actually killing your fundraising.
Your staff works 60-hour weeks. Your opponent's system works 168. Here's why the gap matters more than you think.
Your candidate came home from that fundraiser with 23 business cards. Want to bet how many made it into your CRM?
Every fundraising email waits for candidate approval. Every social post needs their sign-off. What if your content sounded like them, without waiting?
Your candidate promised to send information. Your staff promised to circle back. The donor promised to 'think about it.' Everyone forgot. Here's what that actually cost you.
Ready to build an operation that never sleeps?
SEE THE MACHINE →