My Community OS with AI

It's all about the follow up.

When managing volunteering/community events, things get dropped all the time. We're all volunteers, have many other responsibilities. We over-commit. We forget.

I still drop things, but AI has made it huge impact on my ability to mobilize my community for good. And everytime, its the same pattern, the same framework, the same operating system.

Here is a mocktail of tools that have helped me the most, within the framework of how I see the work.

Operating a Community

In a community the resources are people, time/attention. People want connection, want safety, want opportunities to share their knowledge/expertise, and more. But without some coordination, the impact is much less.

AI doesn't bring anything new to this. But it can lower friction. And in an environment where you have very tiny slices of peoples attention, less friction gives huge impact multiples.

The Personas

You have three main ones - the committee, the volunteers, the participants.

The committee needs to know what to do, what was discussed, what was decided, a roadmap, and a plan. They invest a ton (meetings, setup, etc) and they need it to pay off. Where AI helps, is clarity and alignment. Better notes, better next steps, follow up action items, check lists, etc. So much gets lost without AI, and so much gets recovered with it.

The volunteers need engagement, follow up, scheduling, role assignment etc. They don't have much time, and want to make the most of the day that they signed up. Where AI helps here, is scale. There is a lot of back of forth here. You can "mechanize" everything into spreadsheets, but doing so loses connection with people who are offering their time.

The participants need an awesome event. Overall light touch for communication, announce, RSVPs, day of event, thank you/survey. AI helps somewhat with marketing. Reaching out to targeted people, etc. I dont think I've done a great job here - Day after event is just as important as day off, and its hard to prioritize that.

The Volunteer Funnel

Marketing -> Interested Volunteers -> Engaged Volunteers -> Committed Volunteers -> Delivering Volunteers -> Community Impact

The tools

• Google Form for getting initial interest. API's are pretty solid for making the Forms, and pretty solid for getting responses.
• Email automation via smtp. I found icloud easier to automate than gmail, and that's been working well.
• Text messaging via AppleScript/iMessage works great. For getting responses though, I can read iMessage but I can't read SMS messages.

Tools I don't have

• WhatsApp - I have not gotten this to work. Too risky to tie to my personal account. So much happens here, but it's a black box/manual bridge.
• Partiful/Eventbrite bridge - Actual event management. I wish there was a "events.md" AI native platform 😂

Each of the steps of steps of the funnel is an engagement. Being organized means no one is missed. Reminders to people who haven't responded helps with conversion across the funnel.

Ofcourse, I don't forget to meet/talk with some of the volunteers. Who are they? What are they looking for? Anything I can help with? Any other programs I know they might be interested in? The human part is still there.

The Alignment Funnel

At the micro level - Open Questions -> Agenda -> Meeting -> Decisions, More Questions, Action Items -> Follow Up -> Next meeting

At the macro level - Initiative -> Plan for event/program -> Execution of event/program -> learnings -> Next Plan

In each of these checkpoints, there is a lot to align on, and a lot gets dropped.

Tools

• Markdown Repo - All the "institutional knowledge" about the programs, I keep in github, with markdown. Not Roam, not obsidian, just vanilla markdown, for now. Cloud Code mobile works amazing with it. The right CLAUDE.md file makes entity extraction, todo extraction, profile extraction, follow up extractions a breeze.
• Voice notes -> Otter.ai, and iOS Voice Notes are great for capturing notes. I dont capture recordings in meetings, I do debrief walks, and import the notes. AI then creates follow up items, summary, etc.
• WhatsApp export dump -> Lots of chatter happens in WhatsApp. Export and extract into the repo.
• Retro Questions -> End of event, AI reviews what happened, and shares questions I use to seed a retro.
• "The Checklist" -> End of the retro, end of the event, end of all the conversations, end of all the wins and mistakes, a checklist is born. This seeds the next event, and help bring structure. AI generates the checklist with a holistic view of the data, but I bring it to the committee for alignment.

What's missing

• It's too invasive to do meeting transcripts/voice recordings
• There isn't a great way to "share" the institutional knowledge. I have artifacts, like the checklist, or program description, or retro summary. But I can't say, here is everything I know, have your AI talk to my AI, and help you with your next event.

Conclusion

In this space, of mobilizing a community, there is no Single Source of Truth. There is no hierarchy. There is overarching compensation program. There is no tooling mandate.

I need to meet people where they are, in their channels. Being able to bring a few things together, as hacky as it is, helps a ton. A this feeling that, each event seeds the next one, is a great feeling to have.

Generational Knowledge - This will be the foundation of successful organizations/instituions of the future.

If you have a community you care about, explore some of this.

Actively Kernelling,
Ibrahim