Internal strategy · 90-day plan
A 90-day plan to redesign how Peacemakers works — so two or three people can do the work of ten, without losing the human core of what we do.
The point of view
Most organizations adopt AI by bolting it on top of how they already work — a plugin here, a chatbot there, a faster draft now and then. The gains are real but modest. They look the same as they always did, just slightly faster.
AI-native organizations are different. They redesign their workflows around what AI does well: reading everything, drafting at high velocity, synthesizing across sources, running consistently. The humans in those orgs spend their time on the things AI can't do — presence, discernment, hard conversations, judgment under pressure.
For Peacemakers specifically, this is the opportunity: a three-person team that operates with the back-office capacity of a fifteen-person team, and a course experience that meets people where they are — at 2am, in their kitchen, mid-conflict — without burning out a single facilitator.
"AI does the work that scales. Humans do the work that doesn't."
The plan at a glance
The sequence matters. We connect systems before we automate them, and we automate operations before we put AI in front of participants. Skipping ahead means building on sand.
01
Connect & capture
02
Build & automate
03
Embed in the product
Right now, Peacemakers' data is scattered across QuickBooks, Google Docs, Gmail, the hub, and Seth's head. AI-native starts with making all of it readable in one place. The point of this phase isn't automation yet — it's context. Once Claude can see everything, every later move compounds.
Move 1
Connect the financial backbone.
Install the Claude small-business plugin so QuickBooks is readable in conversation. Income, expenses, donor activity, cash position — all available without exporting a single spreadsheet.
Move 2
Connect the workspace.
Bring Google Drive, Docs, Gmail, and Calendar into Claude's reach. The agenda, supporter tracker, and working docs become live context — not files Seth has to copy in.
Move 3
Choose a donor system of record.
Pick the CRM that will hold relationships for the long haul (Givebutter is a strong fit for a nonprofit of this stage). If not ready, a structured Google Sheet works — what matters is one clean source of donor truth that Claude can reason over.
Move 4
Codify what Claude already knows.
Audit memory and the brand skill. Add what's missing: theory of change, course structure, board member preferences, donor history, common-question answers. This is the org's memory becoming a real asset.
The pattern from the brand skill is the model. Every workflow that runs more than three times a year gets captured once and run thereafter in seconds. The goal of this phase is to take roughly ten hours of recurring weekly work off the team's plate without losing quality or voice.
Move 1
Build the high-leverage skills.
Curriculum content writer (locked to brand voice and the secular-sources rule). Grant application drafter. Donor thank-you writer. Cohort feedback synthesizer. Board prep brief generator. Each one captures a workflow that currently sits in someone's head.
Move 2
Stand up the recurring briefs.
A Monday morning brief covering the week ahead. A monthly impact digest pulling from QuickBooks, surveys, and notes. A pre-board briefing 48 hours before each meeting. A weekly cohort synthesis during course pilots. The org gets a chief of staff that never sleeps and never forgets.
Move 3
Adopt the "AI first draft" default.
Every document, email, application, and plan starts as an AI draft against memory and brand. The team edits up from a draft — never from a blank page. This single habit is responsible for most of the velocity gain.
Move 4
Measure what changed.
End of phase two: write down what used to take a day that now takes an hour, and what used to take an hour that now takes five minutes. This is the receipt — and the case for going further.
This is where most nonprofits stop, and it's the biggest miss. Peacemakers' product is formation — helping people become peacemakers in their actual lives. AI can extend that work into the spaces where a facilitator can't be: the moment of conflict, the 2am reflection, the second language. Done well, this is the difference between a course and a movement.
Move 1
Prototype the course companion.
For Lindsey's first pilot cohort: a chat experience trained on the curriculum that helps participants apply concepts between sessions. Not a replacement for cohort discussion — an extension of it that meets people in the moment they're actually stuck.
Move 2
Build the reflection prompt engine.
Daily or weekly personalized practice prompts, delivered by email, tuned to where each participant is in the curriculum and what they've shared. The kind of accompaniment that used to require a spiritual director.
Move 3
Open the hub to questions.
An "Ask Peacemakers" experience on peacemakers-hub — visitors can ask real questions about peacemaking and get answers grounded in the principles and sources. This becomes a discovery tool, a teaching tool, and a quiet on-ramp into the work.
Move 4
Translate, at minimum, into Spanish.
Peacemaking content has obvious cross-cultural reach. AI-native translation removes the cost barrier that has historically kept small nonprofits monolingual. Pilot with the principle statements and the first course module.
The honest tradeoff
The trap with "AI native" is building so many automations that the organization becomes brittle and people stop thinking. For Peacemakers, where the work is fundamentally about human formation and relationship, this trap would be fatal. The rule has to stay clean.
Drafting. Reading. Synthesizing. Translating. Briefing. Searching. Scheduling. Reconciling. Reporting. Following up. The work that scales linearly with effort and degrades quality when rushed. The work that no one's calling on Peacemakers to do better than anyone else.
Presence. Discernment. Hard conversations. Judgment under pressure. Sitting with someone in pain. Naming what's actually happening in a room. Making the call that no rubric covers. The work that is Peacemakers — and the work participants and donors are actually paying for, whether they say so or not.
If we keep that line clean, AI is a multiplier on what Peacemakers already is. If we blur it, AI quietly replaces the thing that makes Peacemakers worth doing in the first place. The 90-day plan is built to honor the line at every step.
What this unlocks
If we execute this plan, here's what the team should feel by August:
Most weekly admin work — donor follow-up, board prep, financial reporting, content drafting — takes a fraction of the time it does today. The Monday brief, the monthly digest, and the pre-board packet show up on their own. Lindsey's first cohort has a 24/7 companion that extends what she does in the room. The hub answers real questions from visitors who would otherwise have bounced. Spanish-speaking communities can read the principles in their language. And the team's attention is freed up for the work that actually requires them — the conversations, the listening, the judgment calls.
In dollar terms, the equivalent of two or three FTEs of capacity, without the headcount. In mission terms, a movement that can reach further than its budget should allow.
Signs it's working
Three honest signals, at the end of ninety days:
One. The team can name three specific tasks that used to eat hours and now eat minutes — and can show the receipt.
Two. A participant in Lindsey's pilot describes an AI-assisted moment between sessions that they couldn't have had any other way. Not "this was cool." Something like: "I almost gave up on a hard conversation and the companion helped me stay in it."
Three. Seth and Natalie report — without prompting — that their time is going to higher-leverage work than it was in May. Not "I'm less busy." Something like: "I'm spending my Tuesdays on the work I actually started this for."
If we hit all three, the plan worked. If we hit one or two, we adjust. If we hit none, we stop and ask why before doing more.