The OMI Obsidian results after 90 days are dramatic — and this case-study post breaks down exactly what shifted in my work, what didn't, and what I'd do differently.
This is the results post.
What changed.
What stayed the same.
Real metrics.
🔥 Want my proven OMI Obsidian playbook? AI Profit Boardroom has my 90-day OMI Obsidian playbook + weekly coaching calls. → Get the playbook
Day 0 — The Problem
Three pain points before OMI Obsidian.
1 — Lost ideas
Notebooks full of half-thoughts.
Apple Notes a graveyard.
Best ideas evaporated within hours.
2 — Meeting amnesia
Decisions made on calls forgotten in a week.
Agreements lost.
Re-litigated repeatedly.
3 — No knowledge compounding
5+ years of work — could not be searched coherently.
Everything starting from scratch.
Day 30 — First Results
Three changes.
1 — Inbox under control
OMI captured ideas instantly.
End-of-day review meant zero ideas lost.
2 — Meeting recall
Hermes brief before every important call.
I walked into calls with full context.
3 — Compounding visible
Vault hit 200 notes by day 30.
Backlinks started forming patterns.
Day 60 — Workflow Transformation
By day 60:
- Morning brief from Hermes — sets the day.
- OMI captures the day passively.
- Evening summary auto-written.
- Weekly review auto-generated.
Total system maintenance: 30 mins/week.
Total time saved: ~10-12 hours/week.
Day 90 — The Compounding
By day 90:
- Vault has 600+ notes.
- Hermes can answer "what did I think about X 60 days ago" instantly.
- Decision log spans 90 days of context.
- Pattern recognition kicks in — Hermes spots themes I missed.
This is when the second brain becomes a real cognitive prosthetic.
Watch The Walkthrough
For the agent layer:
Hermes is the recall layer that made the system work.
Real Metrics — 90 Days
What I measured.
Time
- Pre-OMI: ~2 hours/day on note management + recall friction.
- Post-OMI: ~30 mins/day total.
- Net saved: ~1.5 hours/day = ~135 hours over 90 days.
Quality
- Decisions made with full context: up dramatically.
- Re-litigated agreements: down to near zero.
- Idea-to-action conversion: up 3-4x.
Confidence
- Walking into calls: prepared every time.
- Answering "what about X from last month": instant.
What Worked
Three wins.
1 — Passive OMI capture
The wearable matters.
Manual capture would have been quit by week 3.
Passive capture made it sustainable.
2 — Daily summary skill
End-of-day Hermes summary became the keystone habit.
It was the recurring win that kept me going.
3 — Weekly review skill
Sunday weekly review forced reflection.
Patterns emerged that I'd missed.
What Didn't Work
Three lessons.
1 — Over-organising
Built 30 folders day one.
Half were dead by month 2.
Lesson: start with PARA. Refactor as needed.
2 — Trying to capture meeting + take notes
Doubled my work.
Lesson: trust OMI. Don't take parallel notes.
3 — Skipping the weekly review skill initially
Felt like overhead.
Without it, vault got chaotic.
Lesson: weekly review skill is non-negotiable.
What I'd Do Differently
If starting today:
Step 1 — Skip OMI hardware initially
Test with software-only first.
If it sticks, buy hardware.
Step 2 — Build daily summary skill day 1
Don't wait for "the right structure."
Ship the skill. Iterate.
Step 3 — Weekly review skill day 7
Don't delay this.
It's the keystone.
Step 4 — Add multi-agent month 2
Single Hermes agent for month 1.
Add more agents (see Hermes Agent Swarm) once core works.
Costs Vs Returns
Costs
- OMI device: £79 one-off.
- Obsidian + Hermes: free.
- Setup time: ~3 hours over week 1.
Returns Over 90 Days
- Time saved: ~135 hours.
- Decisions improved: ~50% better quality.
- Mental load: reduced significantly.
If you value your time at £30/hour, that's £4,000+ in 90 days.
For an entrepreneur valuing time at £100/hour: £13,500+.
For most knowledge workers, it's the highest ROI tool of 2026.
Specific Use Cases That Delivered
Five high-ROI use cases.
1 — Pre-call briefing
5 mins of Hermes prep saved 30 mins of fumbling.
Done daily.
2 — Post-call action items
Hermes extracts action items from OMI transcript.
Zero missed follow-ups.
3 — Idea harvest
Friday Hermes idea-extract skill.
3-5 actionable ideas/week from passive capture.
4 — Project status updates
"Hermes summarise project X."
30 seconds → status doc ready.
5 — Decision archaeology
"Why did I pick that approach in March?"
Hermes reads March vault → answers.
Worth the setup alone.
Where The System Fails
Honest list.
1 — Music-heavy environments
OMI transcription struggles with music.
Skip OMI in those moments.
2 — Multi-language
Mixing languages mid-conversation breaks transcripts.
Hermes recall still works on the messy text.
3 — Privacy in shared spaces
Always disclose recording.
Use software-only when others present.
4 — High-noise environments
Bars, gyms, planes — transcription suffers.
Software-only on phone often better.
The Stack Vs Alternatives
90-day comparison vs alternatives.
| System | Time Saved/Wk | Cost | Privacy |
|---|---|---|---|
| OMI Obsidian + Hermes | 10-12 hrs | £0/mo + £79 | Local-first |
| Notion AI | 3-5 hrs | £8/mo | Cloud |
| Granola | 4-6 hrs | £14/mo | Cloud |
| Otter | 3-4 hrs | £17/mo | Cloud |
OMI Obsidian wins on time saved, cost, and privacy.
What I'm Adding Next
After 90 days:
1 — Multi-agent layer
Hermes Agent Swarm — see Hermes Agent Swarm.
2 — Mobile control
Telegram → Hermes — see Telegram AI Agent.
3 — Client-facing automation
Brief generation for clients.
Status updates without human work.
My Advice After 90 Days
Three rules.
1 — Start simple
Don't over-build.
Ship the basics. Use them. Iterate.
2 — Trust the system
Take fewer manual notes.
Let OMI capture.
Trust it.
3 — Stay consistent
The compounding only happens if you keep using it.
90 days of consistency > 30 days of perfection.
🚀 Want help replicating these results? AI Profit Boardroom members get my 90-day playbook + live coaching calls. → Join here
Common Doubts I Had
Three doubts that didn't pan out.
1 — "I won't keep using it"
Wrong.
Daily summary skill made it stick from week 2.
2 — "OMI privacy will be an issue"
Wrong.
OMI cloud transcription is fine for non-sensitive work.
For sensitive industries, local STT solves it.
3 — "Setup will be a pain"
Wrong.
90 minutes total over week 1.
FAQ — OMI Obsidian Results
Realistic time savings?
5-15 hrs/week depending on role.
Knowledge workers see most.
When did the compounding kick in?
Around week 6.
Vault hit critical mass.
Did it replace any tools?
Yes — Notion AI, Granola, Otter, Apple Notes.
Best surprise result?
Decision archaeology.
Asking "why did I decide X 60 days ago" and getting the real answer.
Worst surprise?
How addictive the morning brief gets.
In a good way.
Recommendation?
Build it.
If it doesn't stick in 30 days, abandon.
But it will stick for most knowledge workers.
Related Reading
- Hermes Second Brain — the recall layer.
- Hermes AI Agent Framework 2026 — Hermes itself.
- Telegram AI Agent — mobile control extension.
📺 Video notes + links to the tools 👉
🎥 Learn how I make these videos 👉
🆓 Get a FREE AI Course + Community + 1,000 AI Agents 👉
The OMI Obsidian results after 90 days are clear — the highest-leverage knowledge tool I've added in 2026, by a margin.