BeroAI · Bero Smart Collar · AI IoT
Turning smart collar data into useful, trustworthy guidance for dog owners.
BeroAI connects a smart collar and companion app to help dog owners understand what may be happening and respond with more confidence. I also designed a separate pre-order site to explain the product and capture early demand.
The product question
BeroAI is an AI IoT pet communication service built around a smart collar and companion app. The service had to turn collar signals into a useful answer: what may be happening, what the owner can do now, and how care can continue over time.
I led product strategy, app UX/UI, AI communication logic, and avatar interaction. In parallel, I designed a pre-order site to explain the product and capture early demand; it supported the launch without becoming part of the app service.
- Reframed BeroAI from a tracking device into an AI communication service.
- Mapped categories for emotion, behavior, and needs with the AI team so interpretation could become usable product logic.
- Structured the app around Understand, Respond, and Bond loops.
| Area | Scope |
|---|---|
| Product | Bero Smart Collar — AI IoT pet communication service built around a smart collar and companion app |
| Role | Product Designer, Product Strategy, Service Design, UX/UI Design |
| Output | App IA and flows, mobile screen system, Emotion-Behavior-Needs mapping, scenario logic, pre-order site and demand-capture flow |
| Status | Pre-order site launched; mobile prototype logic defined |
The question behind the product
Dog owners can see activity, camera footage, or health numbers. The hard part is deciding what a change means and whether it needs a response.
BeroAI therefore had to explain context and next steps without pretending that AI could know a dog's inner state.
- Raw signals had to become guidance that an owner could understand.
- AI interpretations had to communicate likelihood, not certainty.
- The product story had to establish value before asking for a pre-order.
The questions that shaped the product
Before designing screens, I used a Pre-PRD questionnaire to turn the broad idea of an AI pet communicator into explicit product decisions. It covered owner uncertainty, dog-signal taxonomy, explanation logic, privacy, reliability, and responsible engagement.
This was an internal product-definition exercise, not a usability study. Its purpose was to expose assumptions and align the product and AI teams before interface design.
| Decision area | Questions explored | How it shaped the product |
|---|---|---|
| Owner uncertainty | What owners need to know when they are away, walking, or seeing unusual behavior. | Framed BeroAI around practical care moments instead of device novelty. |
| Emotion, behavior, and needs | Which emotional states, behavior labels, needs, and conflict rules the AI should support. | Created the shared mapping for avatar states, need icons, and chat explanations. |
| Narrative and reports | How behavior, emotion, activity, and needs should become diaries, trends, and alerts. | Shifted History and Report from metric dumping to explanations of change. |
| Trust, privacy, and reliability | How to handle consent, AI data use, signal loss, confidence thresholds, and latency. | Set the rule that BeroAI should communicate context and likelihood, not certainty. |
| Care loops and engagement | What Auto Cue, scoring, rewards, and Dog Space should encourage or avoid. | Kept engagement tied to responsible care actions, not shallow gamification. |
Set the promise before designing the interface
For BeroAI to earn trust, I first had to set the boundary of its promise. Rather than claim perfect dog translation, I framed it as a service that interprets patterns, explains possible context, and helps owners respond at a better time.
| Instead of | I positioned BeroAI as |
|---|---|
| A smart collar that tracks activity | A service that turns dog behavior signals into context and care guidance. |
| An AI that "translates" dogs perfectly | An AI communicator that helps owners interpret patterns and respond with better timing. |
| A price-led launch message | A trust-led story that shows why the product matters before asking users to pre-order. |
Three loops: understand, respond, bond
I organized the product around three connected loops: understand the dog's state, respond with an appropriate action, and build the bond through repeated care, training, and rewards.
| Layer | Purpose | Experience output |
|---|---|---|
| Smart collar | Collect behavior, motion, sound, location, and routine signals. | Raw signals and confidence limits |
| AI interpretation and mapping | Translate signal patterns into emotion, behavior, needs, and likely context. | Avatar state, need icon, chat explanation |
| Care surfaces | Make the dog's status and next action visible in the app. | Care Home, Chat Room, History |
| Action systems | Support outdoor safety, activity records, training practice, and repeat engagement. | Walk report, route insight, training journey, missions |
Show the signal, then explain the context
The Chat Room became the explanation layer. Need icons show immediate states such as hunger, potty, or thirst; Chat explains the context, the deviation, and why a response may help.
Each message moved through a scenario: trigger, interpretation, owner action, and resolution. It needed enough reasoning to feel credible without implying that the system knew the dog's exact inner state.
| Scenario | Signal pattern | Owner-facing interpretation |
|---|---|---|
| Unfamiliar noise | Unusual sound, fixed head direction, barking | The dog may be alert because of an unfamiliar sound outside. |
| Activity-based hunger | High activity, mealtime proximity, continued hunger cues | The need icon can show hunger; Chat explains why it may be happening earlier than usual. |
| Separation anxiety | Frozen posture near entrance, micro howling, owner away | The dog may be waiting anxiously and reacting to outside footsteps. |
Answer the first question first
Care Home became the app's primary status surface. It answers the owner's first question: what is happening right now? The answer appears through the avatar, status cards, chat preview, and quick actions.
- The avatar and status cards summarize the current state before exposing raw metrics.
- Chat Preview shows the latest explanation behind a behavior change and opens the full Chat Room when more context is needed.
- Quick Actions lead to a relevant care or training response, while device and account controls remain secondary.
Keep bonding tied to care
Dog Space was designed as the bonding loop. It makes the product feel playful while keeping customization, rewards, and return paths connected to care, walking, training, and repeated visits.
- The room and avatar reflect the dog's recent context and provide a consistent place to return to.
- Customization and rewards stay connected to walks, training, and records rather than becoming a separate game.
- Closing Dog Space returns to Care Home, and blocked states lead to a clear recovery path.
Turn understanding into action
Map and Training gave the service an action layer. Owners could move from understanding a signal to doing something useful outdoors, during practice, or in daily routines.
- Walk insights turn routes into behavior patterns through heatmaps, sniffing signals, stress zones, and reports.
- Safety features surface nearby risks through alerts, privacy zones, and status indicators.
- Training progresses from foundational commands to socialization, alone training, and behavior support.
Explain the product before asking for a pre-order
The pre-order site needed to define a new category, answer trust questions, and collect measurable demand. I led with owner uncertainty and concrete care scenarios before moving into product details and conversion.
| Funnel layer | Design decision | Why it mattered |
|---|---|---|
| Hero and story | Lead with owner uncertainty and the value of understanding the dog. | Make the new product category emotionally legible in the first screen. |
| Lead capture | Collect email, region, optional dog profile fields, and key conversion events. | Support segmentation and measurable demand. |
| FAQ and policy | Address shipping, refunds, subscriptions, hardware fit, warranty, coverage, privacy, and support. | Reduce purchase anxiety for an early hardware and AI service. |
What was ready, and what still needed validation
The pre-order site launched, and the mobile prototype reached a defined product logic. That shows launch readiness, not product-market success.
- Verified: BeroAI's value, service boundaries, and core app structure were defined.
- Verified: The product and AI teams shared an Emotion-Behavior-Needs mapping and scenario logic.
- Verified: The pre-order site launched with lead capture and trust-building FAQ content.
- Not yet verified: interpretation accuracy, real-world care outcomes, retention, and conversion performance.
What this changed in my view of AI design
AI product design is not about making intelligence visible for its own sake. It is about helping people understand what the system is telling them, where its confidence stops, and what they can do next. In BeroAI, that meant connecting hardware signals, AI interpretation, and app experience into one service, while giving the separate pre-order site a clear way to explain that service.