Part 1, Maine, 2018: Campaign Brain Borne From Cycle-After-Cycle of Challenges

Campaign Brain Founder and CEO Nate Levin has spent recent cycles working for Democratic campaigns in all corners of the country, including Maine, Texas, and Arizona. In this series of blog posts, Nate shares how these experiences are informing the solutions we are building at Campaign Brain.

Working as an organizer fighting for Maine Democrats up and down the ballot in 2018, I felt motivated by the potential for artificial intelligence to support our movement at every step, every day.

I had just spent 3 years working for artificial intelligence start-up Fusemachines on question-answering chatbots, computer vision projects for drone delivery of medical supplies to rural villages, and machine learning algorithms to optimize swaths of data.

Now, on the campaign, we were cold-calling potential volunteers and voters one-by-one, tracking data onto paper lists, and trying to optimize our paths through a walk-list with a printed out map.

Eager to do everything I could to contribute to the best interest of the team and ensure we were fighting to elect Maine Democrats up and down the ballot, I accepted the challenge of technological rollback, but was nevertheless enthusiastic about the potential of AI to increase efficiency within campaigns.

In particular, when cutting hundreds of turfs in preparation for the 2018 midterms, it was shocking the required human-effort needed to complete the task, totaling 10 humans for 12 hours a day 3 days in a row. Said another way, we could have knocked another 5,400 doors with that capacity instead of sitting inside cutting turf.

For those who have never cut turf, it is the process of dividing doors that the campaign wants to knock into groups of roughly 25 - 40 doors for a single ‘packet’ that can be knocked on by an organizer or a volunteer during one shift. Done by applying polygons onto a digital map, it is also required to take into account street busyness, dangerous roads, inefficiencies around highways, and density of homes.

While these rule-based limitations are not incredibly straight-forward, if we could build a drone that uses computer vision to guide its way through mountains to locate a rural community, surely we could cut turf using a machine, I thought.

Following victories up and down the ballot in Maine in 2018, I was energized to keep fighting.

With 2020 on the horizon and Trump still in office, there was no time to start building Campaign Brain. There was work to be done immediately, but the seeds of Campaign Brain were already planted.

In the next blog, CEO Nate Levin will share how his time working in Texas shaped Campaign Brain today.