The 90-Day Fundraising Sprint: What Your AI Operation Should Be Doing Right Now

August through October is the final fundraising window before November. Three months to close the money gap, max out your major donor relationships, and run the email frequency that moves the needle.

Campaigns that have their AI operations running correctly going into August have a structural advantage in this window. They're not configuring anything. They're executing.

Campaigns that are still standing up systems in September (still training voice models, still debugging approval workflows, still figuring out their email cadence) will spend the sprint they should be running on setup they should have done in May.

Here's what "fully operational" looks like at the start of the 90-day sprint, and what it takes to get there before August 1.

The State Your Operation Should Be In Right Now

Donor follow-up is triggering within 24 hours of every call session.

Not 48 hours. Not "when the finance director gets to it." Within 24 hours. The follow-up that goes out the morning after a call session lands while the conversation is still warm. The one that goes out three days later lands after the donor has moved on to other things.

If your follow-up system isn't hitting 24-hour velocity consistently, the bottleneck is either the approval workflow or the trigger setup, not the AI. Fix the process, not the tool.

The fundraising email program is running at 2x per week minimum.

Two fundraising emails per week is the floor for an active general election program. One per week leaves donor attention on the table. Less than that and your list goes cold between sends.

If your program is running at 2x, the AI layer should be handling draft generation so the approval step doesn't slow the cadence. Your finance director reviews and approves; she doesn't write from scratch.

The voice model is calibrated to the candidate's current register.

The voice model you trained in January reflected the candidate's communication style in January. By July, the stump speech has evolved, the signature issues have sharpened, the donor conversation has a different rhythm than it did in Q1.

A voice model that's running on January training is generating emails that sound slightly off. Donors who've been in the relationship for six months notice. Recalibrate before August.

Major donor cultivation sequences are active through September.

Major donors don't appear at October fundraisers because someone called them in October. They appear because someone managed the relationship through the summer. The cultivation sequence (research, personal touch, soft ask timing, full ask window) has a 90-day horizon. If cultivation sequences aren't running now, the October major donor results are already set.

The news monitoring source list is current for the general election race.

The source configuration you built for the primary monitored the issues relevant to the primary. The general election race has different opponents, different issue dynamics, and different media markets. Your monitoring setup should reflect where the race is now, not where it was six months ago.

The Configuration Gap

Most campaigns with AI tools deployed aren't running at the state described above. They're running at 60-70% of it; the core is working, but there are gaps that have accumulated since the initial setup.

The voice model hasn't been updated since the primary. The approval workflow is running on the candidate's personal review cycle, which creates a bottleneck every time the candidate is traveling. The email frequency is 1x per week because that's what the approval bandwidth allows, not because it's the right cadence.

These aren't technology problems. They're configuration and process problems. And they're fixable in a few days of focused work before August.

The campaigns that close these gaps before the sprint starts will run August through October at full capacity. The ones that don't will be leaving 15-20% of their AI operation's potential value on the table during the most important three months of the cycle.

What "Not Ready" Actually Costs

The campaigns that don't have this in place by August will feel it in October, but the cost won't be visible in October. It will have accumulated in August and September.

Every week the follow-up velocity is at 48 hours instead of 24: a percentage of those donors are harder to re-engage by the time the follow-up lands. Every week the email program is running at 1x instead of 2x: revenue per email send is lower, list momentum is lower, the relationship with your house file is weaker.

These are small gaps in August. They compound through September. By October, the campaign that built the operation correctly in July is raising meaningfully more per week from the same donor pool.

The window to close the configuration gap is July. Not August. In August, you're running the operation.

If Any of This Isn't Running Yet

There's still time to build it before the sprint starts. The full configuration (voice model, approval workflow, email program setup, donor follow-up trigger, news monitoring, major donor sequences) takes a few weeks to stand up correctly.

Starting in July means the operation is running at capacity by August 1. Starting in August means you're configuring while the sprint is happening, which is worse than not having the system at all; it creates noise and partial execution rather than clean results.

If your operation isn't at the state described here, the conversation to have is now.

Eric Linder is a former California State Assemblyman (2012-2016) and founder of AutomatedTeams, an AI operations consultancy for political campaigns and advocacy organizations.