Asher Elgin Rolls / AI-native growth operatorvol 1.0

[ 05 ]congress

SAIKAT CHAKRABARTI FOR CONGRESS

Six months running growth on a congressional campaign in San Francisco. Instagram went 10,000 to 107,000, the canvassing team's persuasion rate went 35% to 50%, and we came third.

2025-nowGrowth & Outreach LeadPolitical campaign · CA-11, San Francisco

instagram.com/saikatforcongress

Saikat Chakrabarti came third in the primary for California's 11th district, the seat Nancy Pelosi vacated. The audience I was hired to build for him outlived the campaign.

  • whatA three-way jungle primary for Pelosi's old seat. Saikat had worked for AOC and Bernie Sanders, and was an engineer before he was a politician.
  • my partGrowth & Outreach Lead from November 2025. Instagram, volunteer-facing content, canvasser scripting, the call-log layer, and my own voter book.
  • result10,000 to 107,000 followers in six months, third place, and a volunteer network that turned into a PAC still running today.

What I was hired to do

The KPI was straightforward: build Saikat a genuine national audience and real personal credibility, starting from a first-time candidate nobody outside politics knew. Instagram went from 10,000 to 107,000 followers in six months.

What I owned inside that:

  • The content engine, from what got made to when it went out.
  • Scripting and strategy for everything volunteer-facing.
  • Weekly reads of post-level data, moving production toward whatever was turning attention into engagement.
  • Co-branded campaign material with 20-plus local businesses across the district.

Volunteer sign-ups driven partly off Instagram ran into the thousands.

My own book

I canvassed as well as ran the channel. My own book was 100-plus voters a day, at a 15% engagement rate, meaning they picked up and had a real conversation with me, and a 65% persuasion rate on those conversations, meaning they ended in a committed yes.

That number was checked. Seniors re-called 10 to 20 of every 100 yeses I logged, at random, and other volunteers were independently re-contacting the same voters as part of normal turnout work. A padded number would have surfaced inside a week. I held the rate for the length of the campaign.

Reading the call logs

I led a team of ten canvassers, trained them, and A/B tested their scripts door to door. Then I built a Claude-based layer that read the team's call logs and flagged which scripts were failing and where conversations were dropping. The team's average persuasion rate went from 35% to 50% over eight weeks.

The campaign's field program was 1,100-plus volunteers and 185 paid organizers, built by a lot of organizers over months. My slice of it was the social content, the scripting and the strategy that got people through the door, plus the logistics for more than a hundred volunteers across polling locations and canvassing shifts on election day. Our internal polling had Wiener's lead cut from 21 points to 12 in six weeks, with Saikat holding the highest net favourability in the field.

How it ended

Outside spending ran against the campaign through the final stretch. Wiener took 41%. Connie Chan took 30% and advanced to the November runoff. Saikat came third on 18%, about 17,000 votes, with roughly 20,000 people committed to voting for him.

After the primary the campaign converted into a political action committee, Solidarity SF. It backed Connie Chan's runoff and now supports progressive candidates in other races around the country, and I am still working with Saikat's team through it.

the numbers

10,000 → 107,000
Instagram followers in six months, the audience I was hired to build
35% → 50%
the canvassing team's average persuasion rate, lifted over eight weeks by a Claude-based layer that read call logs and flagged failing scripts
65%
persuasion rate on my own book, spot-checked by seniors re-calling logged yeses at random
15%
engagement rate across the 100+ voters I contacted a day
1,100+
volunteers alongside 185 paid organizers, the campaign's field program
18%
third in a three-way jungle primary (Wiener 41%, Chan 30% advanced)