New essay → How I'm building Ouralink in public
Product Case Study

Ouralink

A family operating system for the invisible work of running a home.

In most households one person quietly becomes the operating system — remembering, anticipating, delegating and following up. Ouralink is my attempt to turn that invisible responsibility into something a family can see, own and share.

Role
Founder & Product Owner
Discipline
Discovery · Strategy · AI Product
Category
Family SaaS
Stage
MVP / early validation

This project demonstrates how I approach product problems: understand the user problem → form hypotheses → prioritize tradeoffs → design the experience → use AI where it creates leverage → define measurable outcomes → learn and iterate.

How to read this case study

Ouralink is an MVP-stage product, so nothing here is dressed up as a validated result. Every claim is labelled with what it actually is.

BuiltWorking in the MVP today
In testingCurrently being validated
HypothesisBelieved, not yet proven
ProposedDesigned measurement, no data yet

01 · The problem

The surface problem is logistics.

Every family I spoke to described the same four frictions. They are real — but they are the symptom, not the disease.

Coordinating schedules

School, work, activities and appointments across two calendars.

Remembering tasks

The list nobody wrote down but somebody has to hold.

Household responsibilities

Groceries, laundry, repairs, forms, refills, renewals.

Communicating the work

Explaining what needs doing takes as long as doing it.

The deeper problem

Sharing tasks is not the same as sharing ownership.

A couple can have a shared calendar, a shared grocery list and a shared task app — and one person still has to remember the task exists, decide when it needs to happen, assign it, check whether it happened, follow up when it didn't, and re-plan when something changes.

That loop is the actual work. It never appears on any list, so it never gets divided.

The loop today

  1. 01Remember
  2. 02Plan
  3. 03Delegate
  4. 04Follow up
  5. 05Re-plan

Where Ouralink tries to move it

See
Own
Act
Coordinate
Reflect

02 · The product insight

The problem isn't that families don't have enough productivity tools. The problem is that the responsibility for running the family is invisible, centralized and hard to transfer.

This reframe changed what I was building. If the problem were organization, the answer would be a better list. Because the problem is transferability of responsibility, the product has to make work visible before it can make it shareable — and every MVP decision after this point was measured against that sentence.

03 · Core hypothesis

What I believe, stated so it can be proven wrong.

Hypothesis

If families can make responsibilities visible and assign clear ownership, then cognitive load on the default planner decreases and coordination improves.

Falsified if

  • Responsibilities are created but ownership stays concentrated with one person.
  • Families revert to a private list within four weeks.
  • Assigned responsibilities are completed at the same rate as unassigned ones.

Product principles

  • Solve the coordination problem, not the organization problem.
  • Make invisible work visible before making it shareable.
  • Ownership must be explicit, not implied.
  • Reduce repeated communication, not human communication.
  • Design for participation by the person who is not the planner.

04 · Tradeoffs

What I chose not to build.

Early sketches of Ouralink included health tracking, mood and energy logging, finances, meal planning, wellness dashboards and document storage. Building all of it would have made the hypothesis untestable. Here is how I narrowed it.

FeatureUser valueComplexityMVP decisionReason
Family operations dashboardHighMediumKeepDirectly addresses the core problem by making responsibilities and family operations visible.
Shared tasks & ownershipHighMediumKeepOwnership is the product thesis. Without it, nothing else is worth testing.
Family calendarHighMediumKeepTime coordination is foundational to family operations.
Shopping & household planningHighLowKeepRecurring household responsibilities are a major source of invisible cognitive work.
Financial trackingMediumHighDeprioritizeFamilies already have strong banking and finance products. Important, but it does not solve the mental-load problem.
Family health hubMediumHighDeprioritizeIntroduces privacy and data-sensitivity burden without being necessary to validate the core hypothesis.
Mood / energy trackingLowMediumDeprioritizeInteresting signal, but not essential to proving whether shared ownership improves.
Cycle-synced meal planningMediumHighDeprioritizePotentially differentiated, but too narrow for the initial hypothesis and adds meaningful complexity.
Document storageLowMediumLaterUseful once families live in the product, but it does not move the ownership metric on day one.
Family goalsMediumLowLaterAspirational layer that works best after everyday operations are actually running in the product.

The decision framework I used

Every feature ran through the same six gates.

Core problem
Hypothesis
User value
Complexity
Data / privacy risk
MVP decision

MVP scope was driven by the problem hypothesis, not by how many features could plausibly be added. A feature that scored well on user value but did nothing to test whether families can share ownership was deliberately pushed out of v1.

05 · How I'm building it

From problem → product.

01

Identify the problem

Understand the invisible mental load carried inside busy families — not the surface complaint that everyone is busy.

02

Define the core hypothesis

If families can make responsibilities visible and assign clear ownership, cognitive load drops and coordination improves.

03

Define the MVP

Build only the minimum system needed to test that hypothesis — visibility, ownership, calendar, household planning.

04

Design the family workflow

Model how a need travels from someone's head into a shared, owned, completed responsibility.

05

Validate

Interviews, prototype tests, adoption, ownership distribution and repeat usage — before scaling scope.

The family workflow

How a need becomes shared work.

  1. 01Family need appears
  2. 02Responsibility captured
  3. 03Responsibility becomes visible
  4. 04Ownership assigned
  5. 05Person acts
  6. 06Family sees progress
  7. 07System learns the pattern

How I'd validate it

Evidence before expansion.

User interviewsPrototype testingTask completionFeature adoptionOwnership distributionRepeat usageRetentionQualitative feedback

Proposed validation plan — no results are claimed yet. The sequence matters: qualitative interviews shape the prototype, prototype tests shape the instrumented build, and only behavioral data justifies expanding scope.

06 · AI strategy

Why AI — not AI for the sake of AI.

I did not start with “where can I add AI?”. I started with “where does intelligent automation meaningfully reduce the mental work of running a family?”

Traditional software

The user tells the system what to do.

“Create a task. Set a date. Pick an assignee. Set a repeat rule.” Every field is cognitive work the person carrying the load has to do — again.

The AI opportunity

The system understands context.

Natural languageContextRecurring patternsFamily preferencesPrevious behaviorTask relationshipsTimingOwnership

Turning unstructured family conversation into structured family operations is exactly the work that never gets shared today.

Real AI workflows

Four places AI earns its place.

Trigger

“We need to get Maya's school supplies before Monday and I also need to schedule her dentist appointment.”

Pipeline

  1. 01Natural language input
  2. 02LLM understanding
  3. 03Intent detection
  4. 04Entity extraction
  5. 05Task structuring
  6. 06Confirmation
  7. 07Family system

What the system derives

Task 1
Buy school supplies
Deadline
Before Monday
Task 2
Schedule dentist appointment
Person
Maya
Suggested owner
Speaker
Category
School / Health

Why AI here

Traditional forms make the user do the structuring work. AI removes the structure-entry burden — the family speaks the way they already speak, and the system does the parsing.

Product principle: AI should reduce cognitive load without removing user agency.

AI product principles

The rules I hold the AI to.

AI must solve a real user problem.

AI should remove cognitive work, not add another interface to manage.

Users keep control of decisions that matter.

AI recommendations should be explainable.

Sensitive family information is handled carefully and minimally.

Automation requires confirmation when consequences are meaningful.

AI augments family communication — it does not replace it.

07 · Measurement

How would I know if Ouralink is working?

Not by user count, task count or number of AI interactions. Those measure activity, not whether a family's invisible load actually moved.

North star · proposedNo data yet

Weekly Shared Ownership Actions

The number of family responsibilities that are successfully viewed, assigned and completed by the appropriate family member within a week.

It sits closer to the value proposition than any vanity metric because it can only go up when the work is visible, when ownership actually transfers, and when the owner follows through. A family can create a hundred tasks and score zero.

Visible
Responsibility surfaced to the family
Owned
Assigned to and accepted by a member
Done
Completed by that owner
Proposed

Activation

  • Create or join a family
  • Add another member
  • Create a responsibility
  • Assign ownership
  • Complete a first shared task

Measured as a funnel — where families stall is the real product signal.

Proposed

Engagement

  • Weekly active families
  • Responsibilities created
  • Responsibilities completed
  • Shared ownership actions
  • Repeat planning sessions

Family-level, not user-level — a family is the unit of value.

Proposed

Ownership balance

  • Distribution of recurring responsibilities
  • Share of tasks owned by non-default planner
  • Transfers of ownership over time

Framed as visibility into distribution — never as judging whether a partner is doing enough.

Proposed

Retention

  • Week 1
  • Week 4
  • Month 3
  • Reactivation after a missed week

Family life is seasonal; a missed week is not automatically churn.

Proposed

AI quality

  • Suggestion acceptance rate
  • Generated task accuracy
  • Correction rate
  • Recommendation usefulness
  • Automation completion rate
  • User override rate

AI quality is measured by outcomes and corrections — never by number of AI calls.

08 · Product marketing

How I would position Ouralink.

Feature-led · weak

“An all-in-one family management app with AI-powered task management.”

Describes the software. Says nothing about the person buying it or the feeling they want to stop having.

Problem-led · better

Stop being the person who has to remember everything. Ouralink turns the invisible work of running a family into shared ownership.

Leads with the lived problem, names the villain (invisible work) and states the transformation in the customer's own language.

Target customer

Busy dual-income parents with children, who feel that running the family has quietly become a second job.

Pain point

One person becomes the default planner, reminder system and coordinator for the entire household.

Value proposition

Make family responsibilities visible, assign clear ownership, and reduce the cognitive burden of coordinating everyday life.

Differentiation

Not another calendar. Not another task manager. Not another AI assistant. A shared operating system for family life.

Customer journey

Before and after the shared system.

Before Ouralink

  1. 01Mental load
  2. 02Remember
  3. 03Coordinate
  4. 04Remind
  5. 05Follow up
  6. 06Exhaustion

With Ouralink · intended

  1. 01Capture
  2. 02Make visible
  3. 03Assign ownership
  4. 04Act
  5. 05Coordinate
  6. 06Reflect

09 · Learnings

What building Ouralink taught me.

01

More features don't equal more product value.

A product gets stronger when every feature reinforces one clear problem. Each extra surface I sketched made the core thesis harder to test.

02

The hardest part is deciding what NOT to build.

Cutting financial tracking, health and wellness dashboards was uncomfortable. Prioritization turned out to be the actual product skill.

03

AI should come after understanding the workflow.

If I had started from “where can I add AI?” I would have shipped gimmicks. Starting from the workflow showed where AI genuinely removes work.

04

Automation and ownership are different problems.

Automating a task does not create shared responsibility. A system can be efficient and still leave one person carrying everything.

05

Product metrics should measure the outcome.

If the purpose is reducing invisible load and increasing shared ownership, engagement metrics alone would tell me almost nothing.

06

Building is a series of hypotheses.

The useful discipline is separating what I know from what I believe from what still needs a test in front of real families.

Honest status

What I know · what I assume · what I need to test.

What I know

  • Mental load is unevenly distributed in many households — observed in conversations and widely documented in existing research.
  • Families already use shared calendars and lists and still report a single person coordinating.
  • Sharing a task is not the same as sharing the ownership of remembering it.
  • Existing products optimize individual productivity, not shared household operations.

10 · What's next

What I would build next.

Onboarding that captures one real week

Instead of an empty state, walk a family through describing their actual week and let the system structure it. The first session should already return time.

Ownership rebalancing loop

Turn the delegation workflow into a periodic, opt-in family review with neutral framing and a one-tap transfer.

Instrumented hypothesis testing

Ship the activation funnel and shared-ownership event model first, so every subsequent feature decision has evidence behind it.

Selective depth, not breadth

Revisit deprioritized areas — documents, family goals — only if data shows the core loop retains families first.

Final reflection

Ouralink started because I watched one person hold an entire household in their head. The product work has been less about features and more about answering one question honestly: can responsibility actually be shared, or only relabelled?

I don't have that answer yet — I have a hypothesis, an MVP scoped to test it, an AI strategy pointed at the same thesis, and a measurement plan that would tell me if I'm wrong. That is the part of this project I'd want a hiring team to look at.

Understand the user problem → form hypotheses → prioritize tradeoffs → design the experience → use AI where it creates leverage → define measurable outcomes → learn and iterate.