Home / Case Studies / Smart Journal
Wellness · AIMobile2024Consumer

An AI mood journal that turns daily entries into patterns you can act on.

Smart Journal combines location-aware journaling, AI mood analysis, and a personal insight engine to help users understand what environments, activities, and people make them feel their best. We built the AI processing pipeline, the location tagging system, and the mobile app.

SMART-JOURNAL / HOME Dashboard TODAY Active 78% weekly active rate METRIC 94% ↑ 28% Top item Just now Second item 3 min ago Smart Journal LIVE 2,847 ↑ 18% this week STATUS · ACTIVE 78% weekly active rate 12 weeks Analytics
Client
Smart Journal
Industry
Wellness · AI · Consumer
Surfaces
iOS · Android
Timeline
12 weeks
Outcome
78% weekly active rate
The Problem

Most journaling apps are digital notebooks — they capture text but do nothing with it.

< 15%
Journaling app 30-day retention
~20
Entries before users feel value
60%+
Target weekly active rate

Most journaling apps are digital notebooks — they capture text but do nothing with it. Users journal for a few days, then stop when the habit doesn't feel useful. The problem isn't that people don't want to self-reflect — it's that reflection without feedback loops feels like shouting into a void.

Smart Journal's thesis: an AI that reads your entries, detects patterns in your mood across locations, times, and activities, and surfaces insights you couldn't see yourself — would create a journaling habit worth keeping.

They came to Fusionwave with a clear product vision and a 12-week runway. The AI insight engine had to be genuinely useful from week one — not a novelty that wears off.

What We Designed

Insights that change behaviour.

22 participants
3-week diary study
Insight-first
Core UX pivot
< 2 min
Target daily entry time

We ran a 3-week diary study with 22 participants journaling daily. The finding: users wanted to understand patterns, not re-read old entries. The design pivoted from a chronological journal to an insight-forward interface where patterns surface on the home screen, not buried in a history log.

What We Built

Location + mood + AI patterns.

Flutter
iOS + Android
Fine-tuned NLP
Mood extraction model
Nightly engine
Pattern generation

Flutter handles iOS and Android with CoreLocation/GPS for location tagging. Entries are processed by a fine-tuned sentiment model that extracts mood, energy, and emotional context. A nightly pattern engine runs across the user's entry history to generate weekly insight cards.

Core Features

What ships in the product.

01

Mood-aware entry

Quick daily entry with mood tap, optional voice note, and AI-suggested tags based on content — under 2 minutes daily.

02

Location tagging

Entries are tagged with location context (home, work, gym, park). Mood patterns by place surface automatically.

03

AI insight cards

Weekly cards highlighting mood patterns — 'You feel most energised on Tuesday mornings after gym' — with the entry evidence.

04

Pattern map

Visual map showing mood intensity by location over the past 30 days. Spots where you consistently feel good or bad.

05

Coach suggestions

AI-generated behavioural nudges based on patterns — 'You haven't visited the park in 10 days. Last visit correlated with your best mood week.'

06

Streak and ritual

Daily entry streaks with gentle recovery mechanics. Missing a day prompts a 2-minute voice note reflection — lower barrier than a typed entry.

07

Memory capture

Photo and voice note attachments with automatic speech-to-text. Entries become rich memories, not just text logs.

08

Privacy-first

Entries are E2E encrypted. The AI processing runs on-device for mood extraction — raw text never reaches the server.

09

Export & reports

Monthly PDF mood report with insight summary — shareable with therapists, coaches, or kept as a personal record.

Tech Stack

Production-grade from week one.

Flutter
Core frontend
Node.js
Web layer
CoreML
Data store

We chose the boring, battle-tested options where it mattered and innovated where it gave the product a real edge.

FlutterNode.jsTypeScriptPostgreSQLCoreMLTensorFlow LiteRedisAWSS3GitHub ActionsPostHogSentry
After three weeks, Smart Journal told me I'm always in my best mood on days I walk to work. I'd never made that connection myself. I haven't taken the bus since.
Early user · Smart Journal beta
Challenges Solved

What it took.

  • AI insights that are actually insightful. Early prototypes surfaced obvious patterns ('you feel better on weekends'). We fine-tuned the model on 10,000 annotated journal entries to detect non-obvious correlations — activity sequences, social context, and time-of-day combinations that simple sentiment analysis misses.
  • On-device AI for privacy. Users don't want their journal entries on a server. We fine-tuned a small sentiment model (MobileBERT) that runs entirely on-device for the initial mood extraction — only anonymised mood vectors reach the server for pattern processing.
  • Habit formation mechanics. 30-day retention for journaling apps is typically under 15%. We ran a retention experiment: users shown a personalised insight after their 7th entry had 3x higher 30-day retention than those who weren't. The insight had to come early.
  • Location privacy trust. Location-based features require location access — which users distrust. We built explicit location consent with a clear explanation of what's stored (place category, not exact coordinates), and a 'places I've been' audit log users can delete anytime.
Business Value

What changed.

0%
Weekly active rate (industry avg: 15%)
0x
Retention lift after first AI insight card
0
Average entries before first pattern surfaces
0/5
App Store rating · 2,800+ reviews
App Screens

Key screens, detailed.

A closer look at the core user flows built for this product.

01 / JOURNAL Today Thursday, May 22 CALM MOOD - 7/10 Had a productive morning. Finally finished the report that has been sitting on my desk for weeks. Afternoon slump hit hard AI DETECTED THEMES Productivity Energy AI INSIGHT Your afternoon dips often follow high-output mornings. See pattern Save Entry
02 / PATTERNS Your Patterns Last 30 days - 24 entries MOOD TREND - 30 DAYS Avg 6.8 to 7.6 TOP THEMES Work - 18 entries Exercise - 13 entries Social - 9 entries AI INSIGHT Mood improves 1.4 pts on days you exercise in the morning.
03 / WEEKLY REVIEW Week 21 Review OVERALL MOOD 7.4 /10 +0.6 from last week HIGHLIGHTS Finished major project Consistent theme: achievement Energy dips mid-afternoon Try a 2pm walk this week Social entries boosted mood Plan one social event next week Start Week 22
04 / STREAKS Your Streaks CURRENT STREAK 21 days Personal best: 34 days THIS MONTH M T W T F S S Write Today
Build with us

Building something like Smart Journal?

Smart Journal combines location-aware journaling, AI mood analysis, and a personal insight engine to help users understand what environments, activities, and people make them feel their best. Send us a brief and we'll scope it within 24 hours.

Book a Free Strategy Call →Estimate Project Cost
Call us nowBook a strategy callEmail usChat with our team