Pilot Design Concept · Prototype for Rebuilding Belonging

Coffee isn't
the product.
Belonging is.

Loneliness can trigger damaging mental and physical health spirals. Connection can interrupt them, but creating it intentionally has become harder in a digital-first world.

Double Cup turns a familiar coffee ritual into an invitation anyone can extend: buy two cups, offer one.

Belonging doesn't spread by promotion.
It spreads by invitation.

Double Cup takes that gesture and builds it into community infrastructure—visible membership, neighborhood businesses, and healthcare partnerships working together to make belonging visible, repeatable, and scalable.

Why This Matters

Loneliness has become one of the defining public-health challenges of our time. Across psychology, neuroscience, and public health, the evidence points in the same direction:

People are healthier when they feel connected.

Yet belonging is difficult to prescribe. Long before someone enters therapy, calls a crisis line, or seeks treatment, relationships often begin to shrink. Isolation grows quietly.

You might remember the small acts of kindness that spread during the pandemic—paying for the order of the car behind you at a drive-through. Double Cup builds on that gesture, and adds a choice: carry the extra cup with you and offer it in person, as an invitation to connect—or leave it for someone down the road.

The question this explores: whether thousands of small invitations—each offered one at a time—can add up to something measurable: broader access to community, and to the care that connection makes possible.

01
Part One · The Solution
A community answer to isolation and loneliness — the ritual, the shops, and the people.
A customer with a Double Cup tumbler sits at a café counter beside a sign reading Make It a Double Cup: Round up $2, offer a cup to someone in need, now or later. Member pins and coins sit on the counter while a barista works the pour-over bar beneath a board reading Good Coffee, Brighter Neighbors.
The Double Cup Model
Double Cup member coins — one reading Double Cup Coffee · Member with two cups, the other You Belong Here.

Belonging starts with something people can recognize and a ritual they can repeat.

Small interactions become measurable community infrastructure.

Buy two cups, offer one. The barista hands the extra to you to share in person, or sets it aside for someone down the road. Repeated across neighborhood shops—and supported by membership rewards, health-plan partnerships, and promotion—one small ritual becomes a steady, visible source of human connection.

Illustrative concept — a rendering of the in-shop experience.

01
Who participates?
People · who invite
Coffee shops · welcoming places
Community partners · employers, nonprofits
Health plans · sponsors & evaluators
02
Why they take part?
Connection, and an easy way to give it · People
Loyal regulars, and a role in the neighborhood · Coffee shops
A trusted way to reach the people they serve · Community partners
A measurable channel to hard-to-reach, isolated members · Health plans
03
What does it add up to?
Offer a cup→ A moment of connection→ A coin redeemed, a member counted→ Belonging you can measure
The Double Cup Commitment
“I choose connection over isolation.”

The commitment isn't symbolic. It changes behavior. Members agree to:

  • Notice people
  • Signal openness
  • Offer invitations
  • Welcome others
  • Refer when someone needs more support

Visible symbols make those behaviors easier to recognize and repeat.

Double Cup member goods — member coin, You Belong Here coin, enamel pin, patch, tumbler, cup, cap and beanie
Connection, Made Intentional

Belonging rarely appears all at once. It grows through small, repeated acts of invitation.

  1. Disconnection
  2. Small Invitations
  3. Shared Belonging
  4. Stronger Communities
People
Double Cup Club
Members practice intentional connection.
Places
Double Cup Connector™
Coffee shops create welcoming spaces and know how to invite.
Partners
Double Cup Network
Employers, health plans, and local organizations help connection scale.
Community Model

Coffee Shop Partners & Connectors

Double Cup partners with neighborhood coffee shops already known for more than great coffee. Designated a Connector shop, each becomes the neighborhood’s home for the everyday signals of belonging — recognition people can hold, goods they can wear and carry, and co-branded promotions with Double Cup — backed by Community Connector training that prepares its team to notice, invite, and welcome.

Recognition
A place where everyone belongs
Belonging you can hold. Each piece is a signal that a neighborhood has made room for you, whether that neighborhood has a storefront or not.
La Finca Coffee, a neighborhood café on a tree-lined brick street, displays a Double Cup Connector Shop sidewalk sign reading A Place Where Everyone Belongs. On a café table in front sit a Double Cup mug, a collaboration coffee bag, a Zest x Double Cup collaboration tin, and member coins.
Connector ShopThe neighborhood café, designated and trained.
Illustrative concept
A navy Double Cup Coffee trailer cart parked on a gravel pad beside a rural pasture at dusk, alpacas grazing beyond the fence and a barista at the window. A wooden merch stand holds coffee bags, mugs, and tumblers with a Double Cup tote hanging from its side, and a St. Croix Sanctuary Farms sign bench sits beside it.
Connector CartThe same welcome, on wheels: community gatherings with no café of their own, or at a mission partner like St. Croix Sanctuary Farms ↗.
Illustrative concept
Wear the Invitation
More than merch
Every product is a visible reminder of what a neighborhood is building together — pieces that spark conversation and mark a place to be seen and heard. Spot the symbol, find a place to belong.
Double Cup Coffee apparel lineup on a navy background — a navy cap, cream beanie, navy tee, cream crewneck, and navy hoodie, each carrying the two-cups logo. Double Cup Coffee goods — a travel tumbler, ceramic mug, canvas tote reading More Connection. More Good., bronze Member and You Belong Here coins, and enamel pins and stickers.
Collaborations
Brewed with the neighborhood
Limited co-branded runs — a house blend with a local roaster, a collaboration tin, a seasonal drop — put the shop’s name beside Double Cup’s. Each gives regulars something new to notice, and the shop a real stake in the story it’s telling. Product designs shown are illustrative concepts — example collaborations, not current partnerships.
A Wonderstate Coffee x Double Cup Belonging Blend bag — Driftless, Connected & Safe, medium roast. A Zest Specialty Coffee Roasters x Double Cup collaboration tin, front and back, with a Double Cup Collaboration badge.
Community Connector Training
Belonging you can practice
Being a Connector shop is a practice, not a plaque. The team learns to notice who walked in alone, offer the second cup as a real invitation, and make a newcomer feel expected — small habits that turn a coffee run into a connection.
Notice→Invite→Welcome
"Coffee starts the conversation. Community changes what happens next."
Origin & Opportunity

Inspired by a Simple Tradition

Take two cups.
One is an invitation.

Two Double Cup member coins — one reads Double Cup Coffee · Member with two cups, the other reads You Belong Here.
An invitation you can keep — proof the door stays open.

In twelve-step recovery, one suggested practice is to grab two cups at the coffee pot. One for yourself and one for someone who may need it.

The Double Cup Rule.

The second cup isn’t really about coffee. It’s a simple opportunity to be of service.

The Opportunity

Coffee shops are already trusted neighborhood gathering places. Double Cup helps them become intentional places of connection.

GIVE INVITE GATHER BELONG
A known place to give. A dependable place to connect.
The Idea

Generosity has a pull to it.

Practice small acts of service, in one place, often enough, and that place begins to draw people toward it. People who want to give know where to go. People who could use an invitation know where they’ll find one.

Every second cup adds a little more gravity. Over time, a participating shop becomes more than a place people pass through. It becomes a dependable point of connection, where generosity is visible, belonging feels natural, and everyone is welcome to take part.

You don’t apply to belong.
You belong when you choose to take part.

The Evidence

The Science Is Settled. The Gap Isn’t.

This isn’t only intuition. The link between connection and health has decades of research behind it, across loneliness, mental health, and recovery.

Key Findings
01
Connection is a protective factor.
Across psychology, neuroscience, and public health, strong social relationships consistently reduce the risk of depression, anxiety, substance use, relapse, suicide, and premature mortality.
02
Loneliness changes behavior and biology.
Chronic loneliness is linked to elevated stress, disrupted sleep, impaired cognition, and altered reward pathways. It nudges people toward unhealthy coping and, at the same time, makes reconnecting harder.
03
Mental health and isolation reinforce one another.
People experiencing depression or anxiety often withdraw from others, increasing loneliness and creating a self-reinforcing cycle that can worsen symptoms over time.
04
Recovery happens in relationships.
Treatment still matters, but what makes recovery last is belonging: peer support, a sense of purpose, real relationships, and a community to be part of.
05
Connection deserves to be measured.
The evidence that connection protects health is settled. What's untested is whether a deliberate, everyday intervention can create it at scale and actually move behavioral-health outcomes. That gap, not the science, is what a Double Cup pilot exists to close.
Lit Review
Late 1970s
Environment Matters in Addiction
Bruce K. Alexander · Rat Park Experiments
Rats living in enriched social environments consumed far less morphine than isolated rats, helping shift addiction research toward social and environmental determinants rather than substance exposure alone.
2003–2010
The Biology of Loneliness
John Cacioppo & Louise Hawkley
A series of influential studies demonstrated that loneliness elevates cortisol, increases inflammation, disrupts sleep, impairs cognition, and contributes to depression through measurable biological pathways.
2005
Belonging Protects Against Suicide
Thomas Joiner · Why People Die by Suicide
Joiner's Interpersonal Theory of Suicide identifies two primary drivers of suicidal desire: perceived burdensomeness and thwarted belongingness, emphasizing connection as a critical protective factor.
2015
"The Opposite of Addiction Is Connection"
Johann Hari · Chasing the Scream
Hari synthesized decades of addiction research into a compelling public narrative, arguing that lasting recovery depends on rebuilding purpose, relationships, and community alongside clinical treatment.
2020
Social Connection Should Be Treated as Healthcare
National Academies of Sciences, Engineering & Medicine
A comprehensive review concludes that loneliness is a significant, modifiable health risk. Healthcare systems should routinely identify social isolation and integrate interventions that strengthen social support and community engagement.
2020
Loneliness Predicts Worse Substance Use Outcomes
Ingram et al. · Drug and Alcohol Review
A systematic review found loneliness consistently associated with greater substance use severity, poorer mental health, increased relapse risk, and weaker long-term recovery, supporting social connection as a core treatment target.
2020
Human Connection Is a Health Intervention
Vivek H. Murthy · Together
Drawing on research and clinical experience, Murthy argues that loneliness is an upstream driver of many mental and physical health challenges and that strengthening relationships should be viewed as preventive medicine.
2023
America Faces an Epidemic of Loneliness
U.S. Surgeon General Advisory
Social connection is a fundamental human need. Loneliness and isolation are associated with increased risk of depression, anxiety, substance use, cardiovascular disease, dementia, and premature death. The report calls for rebuilding community as a public health priority.
See the full intervention case in the Impact Report ↗
02
Part Two · The Product
The health-tech intelligence layer — what decides, reads, and measures the ritual.
The Intelligence Layer

Decide where to test,
then learn what worked.

The ritual creates belonging. The product helps a health plan make two accountable decisions. Population Fit weighs population need, delivery feasibility, and missing evidence to recommend a pilot, more discovery, or a decline. Impact OS tracks a running pilot, finds where participation or referrals stall, and helps a reviewer decide what to change.

Rules
Rules calculate.
Every score, threshold, risk estimate, and attribution comes from reproducible arithmetic and classical models that a plan’s analysts can rerun and challenge.
AI
AI explains and drafts.
It turns an approved evidence packet into a plain-language recommendation, summarizes patterns in pilot data, and drafts guidance for a Connector. It works only from what it’s given, and never owns the number.
Human gate
People approve.
Schema and groundedness checks, then a human reviewer, sit between every AI output and any action. The intended data boundary keeps member records out of the model, and the least-autonomous design wins.
Everything shown runs on synthetic data. Each capability is labeled Working prototypeorDesigned next
Module A · Population Fit
Working prototype

Where is a pilot worth testing?

Before a plan spends anything, Population Fit scores its population across six dimensions and returns one of three honest answers: pilot candidate, discovery required, or not a current fit, with the evidence behind each.

InputA plan’s de-identified population and access signals
→
AI taskTurn the scored evidence packet into a plain-language recommendation
→
Reviewed outputA pilot, discovery, or decline memo a person approves

Today a rules-based template writes the recommendation. The Claude connection is built but switched off, and the evaluation harness has to pass before it can be turned on.

  • ScoringWorkingSix-dimension fit score + decision rules. Need alone can never produce a pilot; feasibility, existing programs, and missing data all move the answer.
  • EvidenceWorkingEvidence packet with stable IDs. Supporting, counter, and missing facts, each traceable to a displayed signal.
  • AI seamWorking · templateThe recommendation is written from the packet. A template writes it today; a Claude adapter can take over. Neither can change the verdict or cite evidence that isn’t in the packet.
  • GuardrailWorkingRequest-scope classifier. Refuses individual targeting and clinical or eligibility use, and offers a population-level alternative.
  • EvalWorking20-case golden suite + release gate. Safety gates must hit 100%, and injected model faults flip the gate to Not ready.
3
Honest outcomes, 2 of them “no”
19 / 20
Golden cases pass, 1 kept failing on purpose
100%
Safety-critical gates
Run the assessment → 4 synthetic Minnesota plans
Module B · The Impact OS
Analytics working

Once tested, what happened, and what should change?

An analytics console over a modeled 5,000-member Minnesota cohort tracks participation, referrals, and outcomes with classical models a payer can audit. AI reads and in-the-moment Connector support are designed on top of it.

InputDe-identified pilot aggregates, plus what a Connector chooses to type
→
AI taskSummarize where the funnel stalls; draft codes and Connector guidance
→
Reviewed outputNow / Next / Watch actions and drafts a person accepts, edits, or rejects

The six readouts run today on the modeled cohort. The AI reads and the Connector Companion are designed workflows, shown below with synthetic sample outputs.

  • AI readDesignedPlain-language reads over aggregates. Funnel and siting reads with Now / Next / Watch actions, and a note on what the read did not do.
  • CompanionDesignedThe Connector Companion. Notice, invite, refer, coach: in-the-moment support at the counter, grounded in the public resource directory.
  • CaptureDesignedNote → structured codes. A free-text note becomes Z-codes, navigation codes, and FHIR resources, and a human confirms before anything is submitted.
  • ClassicalWorkingRisk, dormancy, and attribution stay classical. Matched-cohort or stepped-wedge comparisons, so “the AI decided” never enters a payer conversation.
  • MeasureDesignedKill signals and stop criteria, pre-registered. If the effect is indistinguishable from zero, the pilot stops and reports the null result.
6
Live readouts
3
Evidence tiers, consented + aggregate
0
Member records in the AI’s intended inputs
Open the live readouts → Deterministic 5,000-member model
How an output is produced
  1. Aggregate data
    De-identified population & access signals
    Deterministic
  2. Score & rules
    Weights, decision rules, confidence cap
    Deterministic
  3. Evidence packet
    Support, counter, missing, with stable IDs
    Deterministic
  4. AI explains
    Drafts the read or recommendation
    Model
  5. Validate & ground
    Schema, groundedness, decision fidelity
    Guardrail
  6. Human review
    Accept, edit, or reject before any plan action
    Human gate
  7. Measure online
    Leading → lagging indicators, kill signals
    Deterministic

Only step 04 is a model. In Population Fit a rules-based template fills that step today. Switching on Claude means configuring one server-side adapter; the UI, schema, evidence IDs, and release gate stay unchanged, and the API key never ships to the browser.

Example outputs
Population Fit examples come from the working prototype. Impact OS AI panels are designed sample outputs. Every plan, member, and figure is synthetic, and each example shows what the layer produced and what it was not allowed to do.
Population Fit
Impact OS
Recommendation Schema-validated
Pilot candidate 59 / 100 Confidence: high Risk tier: Medium · human approval required

NorthStar Medicare Advantage’s synthetic population shows a credible intersection of social-isolation risk, benefit-navigation friction (71/100), and community feasibility (72/100) near workable partners. Double Cup may warrant a bounded pilot focused on benefit awareness and completed navigation, not a clinical intervention. Confidence is high, but provider-capacity and partner-willingness evidence remains incomplete.

Supporting · 7

  • Social-isolation signal is high (78/100)support_high_isolation_signal
  • Navigation friction 71/100; members aren’t completing navigation despite coveragesupport_high_navigation_friction
  • Community feasibility 72/100support_high_community_feasibility

Counter · 1

  • Member preference for a café-based touch is unproven; signals can align and members still not take up the offercounter_unproven_member_preference

Missing · 4

  • Provider capacity: whether referrals have anywhere to landmissing_provider_capacity
  • Partner-café willingness: no signed commitmentsmissing_partner_willingness
Pilot
6 months · Minneapolis & St. Paul
Primary outcome
Completed navigation referrals from café touchpoints
Expand if
≥ 15% complete a navigation loop across 2 markets
Stop if
< 5% by month 3, any PHI incident, or < 2 active cafés

Next step → Draft a 6-month, two-market pilot charter and take it to human review before any plan action.

How to read these. Population Fit outputs come from the working prototype, with a rules-based template in the AI step. Impact OS AI read and Companion panels are designed sample outputs. All plans, members, and figures are synthetic, and nothing here is a validated product or observed result.

The Case & The Pilot

Measured Belonging. Tested Before Scaling.

Double Cup measures participation and outcomes—not relationships. Reporting is consented, privacy-preserving, and aggregate. The pilot tests one thing: whether this everyday ritual creates belonging worth funding.

90 days
Duration
5–10
Independent coffee shops
250–500
Early members
Does the ritual hold?
Core question

The pilot is where the two parts meet: the ritual is what members experience, and the intelligence layer decides which plan and market go first, then measures whether belonging turned into evidence.

Field Notes ↗
Field Note
Colombian Earthquake Disrupts Major Coffee Access

A M7.4 quake hit the Eje Cafetero and the port of Buenaventura — right where two coffees on my shelf came from.

Read the field note ↗