AI-integrated mobile app
CONSUMER-FACING BRAND · Consumer
A mobile app with on-device and cloud AI features, shipped to the App Store in six weeks.
Kick-off to App Store
6 weeks
6 weeks · 1 PM + 2 engineers + 1 designer
App Store rating at 30 days
4.8★
On-device inference (P50)
~85ms
AI-native at launch
3 features
Before
Where the team was when we picked this up.
- The brand wanted AI in the product but did not want a generic chatbot bolted on.
- Latency had to feel native, not webby.
- Some interactions were too sensitive to send to a third-party API.
What we built
On-device where it counts
Sensitive interactions run a smaller model on-device. The user never sees a spinner and nothing leaves the phone.
Cloud for the heavy lifts
Generation and personalisation hit a hosted model with caching and streaming so the UI never feels stuck.
AI features as product, not paint
Each AI feature has a job. We cut the ones that did not earn their place before launch.
What changed
- Launched on schedule, well-rated, and used.
- Engagement on AI features clears the team’s targets.
- The brand has a roadmap of features instead of one shiny demo.
After
Same team. Same week. Different shape of work.
Stack
React NativeCore MLAnthropic ClaudeOpenAISupabasePostgres
Timeline & team
6 weeks · 1 PM + 2 engineers + 1 designer
More projects
Internal ops platform
A multi-agent ops platform that took ~35 hours of partner busywork out of the week.
~35h/wkManual ops removed
Read RAG copilot
A RAG copilot over 60,000 compliance documents that answers in seconds, with citations.
~94%Answer accuracy on eval set
Read AI memory system
A long-term memory layer for an agent that cut hallucinations and made conversations feel like the user was being remembered.
~70%Drop in hallucination rate
Read Got a workflow like this one?
Book a working session. We will tell you whether this is a four-week build or something bigger, and what it would take to ship it.