# Product Proposal and Initial App Design Created: 2026-08-08 ## Executive Summary This project proposes a next-generation dating and relationship-discovery app built around a simple principle: > **AI driven. Human executed.** The app uses AI to understand each person, identify potentially meaningful connections, design shared experiences, handle coordination, facilitate sensitive communication, and help participants reflect. The humans do the part that matters: they meet, act, notice, feel, and decide. The purpose is not to help users meet more people. It is to help them learn about another person - and themselves - through experiences worth sharing. Instead of asking users to browse profiles, accumulate matches, and sustain text conversations, the app creates opportunities for discovery through action. A first experience might involve delivering meals through a vetted community organization, cooking something together, learning a new skill, exploring an unfamiliar neighborhood, solving a playful challenge, or joining a thoughtfully composed small group. The product can be described as an **AI-guided experience and relationship discovery platform** rather than a conventional dating marketplace. ## The Problem Most dating products are organized around three behaviors: 1. Present a simplified version of oneself in a profile. 2. Evaluate large numbers of other people from limited information. 3. Chat long enough to decide whether meeting feels worthwhile. This structure creates several problems: - Profiles reward self-presentation more than genuine self-knowledge. - Swiping turns people into interchangeable options. - Text chemistry can be misleading and emotionally draining. - Early dates often feel like interviews built around abstract claims. - The number of matches or messages becomes a proxy for progress. - People are asked to predict compatibility before sharing meaningful context. - Difficult but important subjects are either raised awkwardly or avoided. The missing element is not access to more people. It is a better setting in which two people can discover what it is actually like to be together. ## Product Thesis People learn more about compatibility through shared action than through self-description alone. A well-designed experience can reveal how someone listens, responds to uncertainty, treats strangers, collaborates, plays, gives, receives, adapts, and makes another person feel. It can also reveal parts of oneself that do not appear in a questionnaire. The AI's role is to create the conditions for those discoveries without scripting the relationship or replacing human agency. ### The core loop 1. The AI learns about the user privately and over time. 2. It identifies a potentially meaningful person or group. 3. It explains why a shared experience may be worthwhile without over-exposing either person. 4. Participants privately ask the AI what they want to understand. 5. The AI proposes and coordinates an experience suited to the people and the moment. 6. The participants share the experience in real life. 7. The AI invites private reflection and captures discoveries. 8. What it learns informs the next experience, introduction, or personal insight. This is a cycle of **understanding, experience, reflection, and deeper discovery**. ## Product Principles ### 1. Experience before explanation The product should create situations in which qualities can emerge naturally. It should not make every experience feel like an assessment or compatibility test. ### 2. Depth over volume Success is not more matches, more messages, or more time in the app. A small number of worthwhile experiences is preferable to a large funnel of low-context interactions. ### 3. Less chat, more discovery Conversation remains important, but the app does not revolve around an open-ended inbox. Communication should help people feel ready, safe, curious, and able to participate in a real experience. ### 4. AI as guide, not substitute The AI can understand, suggest, coordinate, translate, and reflect. It must not impersonate users, fabricate feelings, or become the relationship itself. ### 5. Mystery with consent The app should preserve curiosity through gradual discovery, but never by withholding safety-critical information or misleading participants. Mystery is not secrecy. Disclosure must be progressive, mutual, and controlled by the person whose information is involved. ### 6. Every experience has intrinsic value The activity should be worthwhile even if no romantic relationship develops. Participants should be able to leave thinking, "I am glad I did that," not merely, "That match failed." ### 7. Learning goes both ways The app helps a user learn about another person and notice patterns in themselves: when they feel open, guarded, energized, generous, playful, or understood. ### 8. Human agency remains final The AI may make a recommendation, but users choose what to reveal, whom to meet, which experience to accept, and whether a connection continues. ## The Experience as the Primary Product Object Traditional dating applications are organized around profiles and message threads. This app is organized around experience invitations and the discoveries that follow them. An experience has: - A reason it may suit these participants - An activity with value of its own - A discovery intention that remains subtle rather than diagnostic - Time, location, cost, travel, and accessibility constraints - Comfort, energy, and social-intensity settings - Safety and consent requirements - A clear logistical plan and fallback - Optional prompts that support natural interaction - A private reflection afterward ### Experience families #### Service - Deliver meals through a vetted community organization - Assemble care packages - Participate in an organized neighborhood cleanup - Help at an established animal shelter Service experiences involving children, elderly people, or other vulnerable populations must occur through vetted organizations with their required screening and supervision. The app should never arrange informal access to vulnerable people. #### Make - Cook a meal from a shared set of ingredients - Take a pottery, printmaking, or flower-arranging class - Build something small for a community space - Create a short photo story together #### Explore - Follow a curated neighborhood trail with optional prompts - Visit a museum and independently choose one object to show the other - Explore a market and select something for a shared meal - Take a low-pressure nature walk with a small observation challenge #### Learn - Attend a beginner class neither person has tried - Learn a simple dance, craft, or outdoor skill - Join a guided tasting or cultural workshop - Teach each other a small skill #### Play - Complete a cooperative game or puzzle - Try a playful scavenger hunt - Join a low-stakes team activity - Make choices in a branching local adventure #### Contribute as a group - Join four to eight compatible people for a volunteer or creative activity - Work in small rotating pairs during a structured event - Share a meal or reflection after completing a common task Small groups can reduce pressure, improve safety, and reveal social qualities that a one-on-one interview-style date may conceal. ## Initial User Journey ### 1. Private orientation The user begins with a conversation rather than a public-profile form. The AI learns about: - Values and relationship intentions - Curiosities and interests - Social energy and preferred pace - Activities that feel nourishing or draining - Practical constraints and accessibility needs - Boundaries, deal-breakers, and safety preferences - How the user wants information shared - What the user hopes to understand about themselves The user can review and correct what the AI believes it knows. ### 2. An invitation, not a match card Instead of presenting an endless catalog, the app offers a considered invitation: > "There is someone I think you may enjoy discovering through a shared experience. You both care about community, though you seem to express that care differently. I think the contrast may be interesting." The initial reveal provides enough information for informed consent without attempting to summarize the person completely. Additional details can be revealed at mutually agreed stages. ### 3. Private questions for the AI Before accepting, each person may ask the AI questions such as: - "Why do you think this could be worthwhile?" - "What kind of experience would suit us?" - "Is there anything important I should know beforehand?" - "How does this person tend to approach new situations?" - "I want to ask about children, sobriety, religion, or politics. Can you help?" - "What information about me would you share if they asked the same thing?" The AI answers only from information the other person has explicitly allowed it to share in that context. When permission is missing, it can ask the other person privately or help the user formulate a respectful direct question. ### 4. Mutual opt-in Both people accept the introduction and the level of information being shared. Declining is private, simple, and does not require a defense. ### 5. Experience proposal The AI presents a small set of options, ideally no more than three. Each option explains: - What participants would do - Why it may suit them - Duration and approximate cost - Energy, conversation, and physical-activity levels - Location, travel, and accessibility details - What has already been arranged ### 6. Coordination The AI gathers availability and constraints separately, proposes a time, and assists with reservations or organization enrollment. No purchase or booking is completed without explicit confirmation. The AI handles logistical conversation so the participants do not have to manufacture rapport through scheduling messages. ### 7. The real-world experience The app recedes. It may provide an itinerary, contact information, optional prompts, and a safety check-in, but it should not demand continual attention. ### 8. Private reflection Afterward, the AI invites a short, adaptive reflection: - What surprised you? - When did you feel most like yourself? - What did you appreciate about the other person? - Was there a moment you became more curious or less comfortable? - Is there something you want to understand next? - Would another experience feel welcome? These responses remain private unless the user deliberately shares them. ### 9. The next chapter If both people are interested, the AI proposes a next experience based on what emerged. A connection becomes an unfolding shared journey, not merely a persistent chat thread. ## Initial App Design The interface should feel calm, intentional, human, and slightly intriguing. It should avoid the visual grammar of catalogs, social feeds, and gamified matching. ### Primary navigation #### Today The home screen presents one meaningful next action: - Continue the private orientation - Consider an invitation - Choose an experience - Confirm logistics - Prepare for an upcoming experience - Reflect on something recently shared There is no infinite feed. #### Guide The private AI relationship guide. Users can ask questions, discuss concerns, correct the AI's understanding, explore difficult topics, and decide what may be shared. The Guide should clearly distinguish: - What the user explicitly said - What the AI tentatively inferred - What is private - What is currently shareable #### Experiences Three simple views: - Upcoming - Invitations - Past experiences Past experiences can include a shared artifact - perhaps a photo, place, achievement, or short mutually approved memory - without becoming a public social feed. #### Discoveries A private, evolving field journal containing: - Things the user has learned about themselves - Patterns the AI has noticed and the user has confirmed - Questions the user is carrying - Meaningful moments the user chose to preserve - Qualities they have come to value through experience The user can edit or reject every AI-generated interpretation. #### Connections Each connection is represented by a journey rather than a profile page. It can show: - What the other person has chosen to share - Experiences completed together - Mutually preserved moments - Open questions - The next proposed chapter - A prominent "Ask the Guide" action Direct messaging, if available, is secondary and bounded. It should not be the default mechanism for keeping the connection alive. ### Key screen: Experience invitation The initial card should contain: - A human introduction with limited, consented detail - Why the AI believes discovery may be worthwhile - The kind of experience being considered - Essential safety and compatibility information - Controls to ask privately, learn a little more, accept, or decline It should not display a compatibility score. The AI should explain its reasoning in natural language and acknowledge uncertainty. ### Key screen: Experience plan The plan should make the real-world activity feel tangible: - Activity title and short narrative - Date, time, place, duration, and cost - What to bring or wear - Accessibility and dietary information - Reservation or partner-organization status - Travel plan - Safety options - Weather or venue fallback - Confirm, request a change, or decline ### Key screen: Reflection Reflection should feel conversational rather than like a survey. The AI may ask one strong question and follow the user's response. The user may mark an insight as: - Private - Helpful for future curation - Shareable with this person - Something they want help discussing ## AI System Responsibilities ### Personal understanding Maintain an evolving model of the user that separates facts, preferences, boundaries, hypotheses, and changing context. Every inference should carry uncertainty and be correctable. ### Connection curation Identify people who may create a meaningful learning opportunity together. Compatibility matters, but productive difference, timing, readiness, and experience fit may also matter. The system should avoid treating people as deterministic personality types. ### Experience design Choose or compose experiences using: - Shared and complementary interests - Desired level of novelty - Social and physical energy - Values participants may enjoy expressing - What each person is curious to learn - Safety, budget, location, schedule, and accessibility - The current stage of the connection ### Consent-aware mediation Answer questions about another person only within that person's permission settings. The AI may relay a question, request permission, coach a conversation, or say that it cannot answer. ### Logistics Coordinate schedules, venues, reservations, partner organizations, reminders, changes, payments where appropriate, and fallback plans. All consequential actions require user confirmation. ### Reflection and learning Help users name discoveries without declaring what they feel or making clinical judgments. The system should offer interpretations tentatively and allow users to correct or delete them. ## Privacy, Consent, and Safety Model The product depends on unusually high trust. Information should be divided into clear layers: | Layer | Purpose | Access | | --- | --- | --- | | Private vault | Personal history, fears, reflections, and sensitive context | User and authorized AI processing only | | Curation signals | Data used to identify people and suitable experiences | Matching and planning systems only | | Shareable self | Information the user permits the AI to communicate | Potential connections under defined conditions | | Contextual permission | Information allowed for one person, group, or stage | Only the specified recipient and context | | Shared memory | Moments both people agree to preserve | Participants in that connection | Important safeguards: - Users can preview what the AI would say about them. - Sensitive inferred traits are never disclosed as facts. - The AI does not impersonate a participant or claim feelings on their behalf. - Identity verification and strong reporting/blocking controls are foundational. - Early experiences default to public places or vetted partner organizations. - Users can share an itinerary with a trusted contact and schedule check-ins. - Participants can leave or decline without having to negotiate through the other person. - The system should not infer consent from prior participation. - Safety-critical information is not hidden in the name of mystery. - Data deletion, correction, retention, and model-training choices must be understandable and accessible. ## What the Product Deliberately Avoids - Infinite swipe feeds - Public popularity signals - Compatibility percentages presented as truth - Engagement mechanics built to delay real-world action - AI-written flirting presented as the user's own words - Unbounded disclosure of private or inferred information - Treating every experience as a covert personality test - Forcing a romantic outcome for an otherwise worthwhile encounter - Measuring success primarily through daily active usage or message volume ## Initial MVP The MVP should test whether AI-curated shared experiences produce meaningful discovery and whether users trust the AI enough to participate. ### Suggested scope - One geographically concentrated, invitation-based cohort - Private conversational onboarding - User-reviewable AI understanding and sharing controls - A limited set of high-quality experience templates - Curated one-on-one invitations - At least one small-group experience format - Private AI-mediated questions before an experience - Availability collection and plan confirmation - Human-assisted reservations and partner coordination behind the scenes - Simple itinerary and safety check-in tools - Private post-experience reflection - Mutual choice about whether to continue In the earliest version, operations can remain human-in-the-loop. The AI can generate recommendations and coordination plans while a small operations team verifies venues, community partners, reservations, and exceptions. This reduces risk and creates the data needed to automate responsibly later. ### What the MVP does not need - A large national user base - Thousands of activity options - Fully autonomous purchasing or reservations - Complex social feeds - Extensive direct messaging - A definitive compatibility algorithm ## Measures of Success Primary measures should reflect the quality of discovery: - Percentage of accepted experiences that occur - "I am glad I shared this experience" response - "I learned something meaningful about this person" response - "I learned something useful about myself" response - Desire for another shared experience - Participants feeling accurately understood by the AI - Trust in the AI's privacy and mediation behavior - Safety incidents, cancellations, and uncomfortable-surprise reports - Quality of group composition and experience-provider reliability Useful anti-metrics: - Match count - Messages per user - Time spent scrolling - Daily engagement without meaningful progress ## Business Model Direction The incentives should align with successful real-world experiences rather than attention capture. Plausible models include: - Membership for ongoing curation and concierge support - Per-experience coordination or booking fees - A hybrid membership with included and premium experiences - Partnerships with vetted venues, instructors, and community organizations, disclosed transparently Advertising and engagement-maximizing mechanics would risk undermining trust. ## Key Product Questions - Who is the best first community: a particular city, life stage, interest community, or relationship intention? - How much information should be revealed before mutual opt-in? - Should photos appear immediately, progressively, or according to user preference? - How should direct messaging be bounded without making users feel controlled? - Which experiences reliably create discovery while remaining enjoyable and accessible? - When should an experience be one-on-one versus group-based? - What is the right balance between algorithmic curation and human experience design? - How should the system handle asymmetric interest after an experience? - What language makes AI mediation feel discreet and supportive rather than invasive? - Which sensitive topics may the AI mediate, and which should remain direct human conversations? ## Positioning ### Short description An AI-guided platform that helps people discover one another through meaningful shared experiences. ### Product promise > Less chat. More discovery. ### Supporting line > Do not just describe who you are. Discover who you are together. ### Strategic distinction Conventional dating apps optimize looking and talking. This product designs experiences through which people can learn who they are together. ## North Star The ideal outcome is not that a user spends more time in the app or meets the largest number of people. It is that they can say: > "That experience helped me see this person more clearly, and I learned something real about myself in the process."