JOURNAL
From Raw Notes to Published Post: The Workflow That Doesn't Feel Like Work
2 — The Voice Problem
6 May 2026 · 7 min read
Every travel blogger has a backlog. Not because they don't want to write — but because the distance between "I have notes from this trip" and "I have a published post" is large enough that it keeps getting deferred.
The old workflow for a typical 1500-word travel post looks like this:
OLD WORKFLOW
═══════════════════════════════════════════════════════════════
1. Find your notes (scattered across Notes app, WhatsApp, photos)
Time: 15-30 min
2. Organise the notes into something usable
Time: 30-45 min
3. Open a document and stare at the blank page
Time: variable (sometimes hours, sometimes you close the tab)
4. Write a first draft
Time: 2-3 hours
5. Edit for structure, voice, readability
Time: 1-2 hours
6. Find and add internal links, SEO metadata, images
Time: 30-45 min
7. Push to CMS, format, preview, fix formatting
Time: 30-45 min
─────────────────────────────────────────────────────────────
TOTAL: 5-7 hours, often spread across multiple sessions
Because of the spread, the backlog grows faster than you clear it
The Roami pipeline doesn't eliminate this work. It restructures it so that the heaviest parts happen at the point of lowest friction — in the moment, on the road — rather than weeks later at a desk.
The full pipeline
ROAMI PIPELINE — OVERVIEW
═══════════════════════════════════════════════════════════════
ON THE ROAD AT THE DESK
─────────── ───────────
[TRIP HAPPENS]
│
▼
[LOG MOMENTS] ──────────────────────────────────
Location + text │
30 sec per entry │
Offline-capable │
│
│ │
▼ │
[ENTRIES SYNC] │
Auto on reconnect │
│ │
└────────────────────────────────────────────────► │
│
[SELECT TRIP + START DRAFT] │
Choose which trip to write about │
Pick post type │
│ │
▼ │
[AGENT READS] │
Your logs for this trip │
Your published posts (voice) │
Your import archive (if any) │
│ │
▼ │
[SCAFFOLD] │
Proposed structure │
Section headings │
Key points per section │
You approve or edit │
│ │
▼ │
[FIRST DRAFT] │
Full post in your voice │
Flags thin sections │
Preserves your phrasing │
│ │
▼ │
[YOU EDIT] │
~20-45 min depending on logs │
│ │
▼ │
[PUBLISH] │
Direct to your CMS │
Or export to WordPress/Ghost │
│
──────────────────────────────────────────────────────────
Step by step, with actual times
Step 1: Log during the trip
Time investment: 1-3 minutes per day while travelling
This is the only step that happens in real time. Everything else happens later at a desk.
The logging is intentionally low-friction — one sentence minimum, location auto-tagged, works offline. The discipline required is: log when something happens that you'd regret not writing down. Not every hour. Not on a schedule. When the moment calls for it.
For a 10-day trip with moderate logging discipline, expect 30-50 entries. That's 30-50 moments with enough detail to reconstruct the trip.
Step 2: Select trip and decide what to write
Time: 5 minutes
Back at a desk, open the archive and look at the trip. The entries tell you what's there. Usually the angle for the post is obvious from reading them — either there's a clear narrative thread running through the logs, or there's a specific destination or day that has enough material for a standalone piece.
For a 10-day Ladakh trip with 44 entries, the decision was: one travelogue (the return narrative — "We Went Back to Ladakh") and one practical guide (the evergreen reference piece). The travelogue came from entries that were emotionally loaded. The guide came from entries that were practically specific.
Step 3: Scaffold review
Time: 10 minutes
Before writing the full draft, the agent proposes a structure — section headings, key points per section, estimated length. This is the moment to redirect if the angle is wrong or a section is missing.
The scaffold for the Ladakh travelogue looked like:
SCAFFOLD — "We Went Back to Ladakh. It Was Still True."
══════════════════════════════════════════════════════════
Opening: The flight into Leh (specific image from logs — mountains 40 min before landing)
Why return: The 2016 trip vs. the 2024 trip — what was different this time
The road: Srinagar route vs. flying — what each one is for
Nubra Valley: The section with the most log density — anchor section
Pangong Tso: The comparison post (same view, different decade)
Altitude: Honest section — what happened, what to expect
The thing that was still true: Closing emotional beat
Practical notes: Brief — links to the guide for detail
If the scaffold is wrong, you tell the agent and it revises. This is much faster than discovering the structure is wrong after the full draft is written.
Step 4: First draft
Time: 5 minutes (for the agent), then you review
The draft comes back as a full post. It uses the logs as raw material and your published posts as voice reference. The sections with the most log density are the strongest; sections where the logs were thin get flagged.
For the Ladakh travelogue, the Nubra Valley section was strong — 12 entries from those three days. The Pangong section was thinner — four entries, all factual. The agent flagged the Pangong section with a note: "Limited personal material — consider adding a memory or removing the emotional framing."
That flag is useful. It tells you exactly where you need to do the work.
Step 5: Editing
Time: 20-45 minutes
This is the step that varies most. The editing is not rewriting — it's:
- Adding the personal details the logs hinted at but didn't capture fully
- Cutting sections that the agent over-explained
- Fixing a sentence that's technically in your style but lands wrong
- Adding internal links and a few external references
- Adjusting the opening if it didn't land right
The Koh Lanta post needed about 20 minutes of editing — the logs were specific and the voice was clean. The Three Trips to Thailand overview needed closer to 45 — the emotional arc needed more work because the logs from three trips had to be synthesised into a single narrative.
Step 6: Publish
Time: 5-10 minutes
From the Roami editor, one click to push to your CMS. If you're on WordPress or Ghost, it goes straight there. If you're hosting your own blog (like WanderingBong), the API handles the transfer.
After pushing: add featured image, check the slug, set the meta description, tag the post. These are five minutes if your image library is organised.
What the pipeline changes
The old workflow has one structural problem: the heaviest work happens furthest from the trip. By the time you sit down to write, the memories are fading and the motivation has dropped with them.
The pipeline inverts this. The hardest cognitive work — deciding what mattered, capturing the specific details — happens during the trip, in 30-second increments, at peak engagement. The desk work is largely editorial: reviewing, shaping, filling in what the agent flagged.
EFFORT DISTRIBUTION COMPARISON
OLD WORKFLOW ROAMI PIPELINE
──────────── ──────────────
ON THE ROAD: ░ (minimal — a few photos) ████░ (logging — the key work)
AT THE DESK: ████████████████████████ ███████░ (editing — focused)
(5-7 hours, often deferred) (30-45 min, same week)
The total effort is not dramatically less. The distribution is completely different — and the distribution matters, because deferred work tends to become abandoned work.
The honest failure mode
The pipeline fails when the logs are too thin. If you spent 10 days in a place and logged three entries — all factual, none personal — the draft will be competent and generic. The agent can't invent the texture it wasn't given.
This is not a product failure. It's a reminder that the pipeline is an amplifier. Good logs produce good drafts. Thin logs produce thin drafts.
Which is why the logging habit is the foundation. The pipeline is only as good as the material that goes into it.
Join the publish tier waitlist →
Related: Why Your AI Travel Writing Doesn't Sound Like You · The Context Problem · The Thailand Case Study
Ready to log your next trip?
Join the waitlist →