With a dominant direct-to-consumer presence and a loyal customer base, the company had built its reputation on premium audio hardware. Traditionally a hardware business, Raycon saw a massive opportunity to build recurring revenue by entering the smart device market. The goal was ambitious: create an AI Notetaker, a screen-less, magnetic device designed to be a personal AI memory bank. Raycon had deep manufacturing expertise, but needed a software partner to validate market demand and build the digital infrastructure, so they turned to NineTwoThree.
Validated market demand through surveys with 2,000+ respondents and live prototype testing.
Built natively on both platforms so background audio transfers stay stable through calls and locked screens.
Conversations turn into actionable briefs automatically, with auto-tagging that keeps recordings organized.
An invisible sync protocol detects the device and moves large audio files over Bluetooth without user action.
Moving from hardware sales to a daily-use software product meant solving three problems at once.
Historically, Raycon's relationship with its customers ended at the point of sale. The challenge was to evolve from one-time hardware transactions into a brand that engages users daily through a high-value mobile app.
The product's core value depended on reliable Bluetooth transfer of large audio files from a physical device to a mobile app, running in the background. A phone call, a locked screen, a brief interruption: any of it could corrupt the transfer and destroy the user experience.
The AI note-taking market is crowded with tools that overwhelm users with features and jargon. Raycon's customers expect simplicity. Delivering sophisticated transcription and summarization without the cognitive load of competitor apps was a real design and engineering challenge.
Working as an embedded R&D partner, we replaced assumptions with data and built a workflow that prioritized speed, stability, and simplicity.
We validated the direction through a strategic workshop and two survey rounds, then ran customer interviews and live prototype testing. Subscription analysis, financial modeling and an NPI readiness checklist guided the project end to end.
2,000+
survey respondents across two validation rounds
We built natively for both iOS and Android with an invisible sync protocol that automatically detects the hardware and initiates secure transfers. If a call comes in or the screen locks mid-transfer, the system continues in the background with no data lost.
0
user actions required to resume interrupted transfers
We focused on a minimal design where generative AI automatically summarizes conversations into actionable briefs. A "Topics" system auto-tags recordings so the app stays useful over time without requiring constant user input.
80%
confidence threshold for auto-tagging recordings
A production-ready native app delivered in 22 weeks, protecting the critical launch window. The timeline covered validation, engineering, and deployment across both iOS and Android.
A scalable SaaS ecosystem with tiered monthly and annual subscriptions, built to offset cloud processing costs and grow with the product over time.
A workflow that outperforms competitors on ease of use, built around automatic actions that match how Raycon's customers actually think about their recordings.
1:1 hardware-to-account binding and remote cryptographic wiping ensure sensitive voice data stays exclusively in the hands of its owner, meeting enterprise security standards.
Build a new platform, or add AI capabilities to existing software, we deliver solutions that work in real operational environments.
Meet with founders Andrew Amann & Pavel Kirillov
Custom AI implementation roadmap
No sales pressure
