← Learning log
Live dashboard · Updated 25 October 2026
Building Bodhi, in public
I'm building an AI chief-of-staff — a small team of agents that researches the news at 6am, drafts my posts in my voice, and briefs me before coffee. This page is its public report card, updated as the build progresses. The deep working files stay private; what you see here is what I'm comfortable sharing.
The crew
What's running right now
Four scheduled agents, each with one job. If something fails twice, it tells me instead of failing silently.
| Agent | Schedule | Job | Status |
|---|---|---|---|
| Research agent | 6:00am daily | Reads primary sources on AI search, writes the day's intel brief | Live |
| Morning brief | 7:30am daily | Top three priorities, delivered to my phone | Live |
| Content approval | 9:00am daily | Drafts platform-native posts, waits for my go-ahead | Live |
| Evening wrap | 8:30pm daily | Scoreboard and lessons from the day | Live |
| Learning cycle | Every 48h | Pulls post metrics, updates what works per platform | This week |
| Pilot outreach | This week | Deep-researched introductions to Australian ecommerce agencies | Queued |
Build log
Week by week
The full write-ups live in the learning log. This is the trail so far.
On the board
What's next
| Next | Why it matters | When |
|---|---|---|
| First learning cycle | 48-hour engagement data feeds back into how every post gets drafted | Tuesday |
| Publishing flow | Approved posts go live here without manual steps | This week |
| First AI-visibility pilots | Free audits for Australian brands, results published openly | November |
| Voice interface | The JARVIS bit — talk to the build, get answers out loud | Later |
Follow the build.
New log entries land here as the system grows. Questions about the build, or an AI-visibility pilot for your brand — that's Centile.