Proof of concept:
Mentor Network

Reframing a matching-app idea around the harder problem of helping trust grow safely over time.

The situation

Young adults leaving foster care can lose the adults they depend on almost overnight. A digital matching experience initially seemed like a possible way to connect young people with screened adults who wanted to help.

But matching vulnerable young adults with unfamiliar adults creates a different kind of problem: the product has to support connection without creating pressure, exploitation, dependency, or a false sense of safety.

I used one week to explore the problem, define early safety rules, and build a concept that could make those assumptions concrete enough to test.

Why it was messy

More choice could create more risk.

A typical marketplace model would optimize for discovery and fast matching. In this context, those same mechanics could expose young adults to unwanted attention, encourage premature trust, and place too much responsibility on them to judge adults they do not know.

Signal

The real problem was not matching people

The research and program models I reviewed suggested that young adults need relationships that are voluntary and steady, while also needing protection from exploitation, pressure, and sudden abandonment.

That changed the problem for me. Instead of asking how to create better matches, I started asking how a system might let trust develop gradually, visibly, and with room to leave.

Signal 1

Choice still matters

A safety model should not remove the young adult’s agency or make relationships feel assigned.

Signal 2

One relationship is fragile

Relying on one adult creates a single point of failure if that person becomes unavailable or the relationship does not work.

Signal 3

Trust needs time

Early interactions need more visibility and structure than established relationships do.

Decision

Replace browsing and matching with requests for specific help

Instead of showing young adults a catalog of adult profiles, the concept lets them ask a screened network for specific kinds of support. Relationships can then grow out of real interactions rather than a profile-based match.

The model also uses a small support circle rather than placing all responsibility on one mentor.

Tradeoff

Choose safety over speed

I deliberately removed mechanics that would make the product feel faster or more familiar: no open profile browsing, instant matches, or unrestricted early messaging.

  • Young adults ask for help instead of browsing adult profiles.
  • A coordinator can remain visible during early relationship phases.
  • Relationships grow through stages rather than becoming fully open immediately.
  • A support circle adds redundancy around one primary trusted adult.
  • Reporting concerns and ending contact stay visible and easy to find.

Direction

Make the safety model visible in the product

The prototype turns the trust model into something a young adult could actually react to. The four primary areas focus on orientation, transparent requests, relationship structure, and communication boundaries.

01: Overview

See the full support picture at a glance

The dashboard brings active requests, the current relationship phase, unread messages, upcoming activities, and the young adult’s support circle into one view. Safety actions remain visible rather than being buried in settings.

Compass overview dashboard showing active requests, support-circle size, unread messages, current phase, an upcoming group meal, and safety controls
The overview keeps orientation, upcoming activity, relationship status, and safety controls in one place.

02: Requests

Make asking for help transparent and trackable

Young adults can see requests awaiting coordinator review, currently active, or already completed. Each request shows the people involved and what happens next.

Compass My Requests page showing pending coordinator review, an active group meal request, and a completed transportation request
Clear statuses make the coordinator’s role visible and reduce uncertainty about what happens after asking for help.

03: Support circle

Build redundancy around one trusted anchor

The circle distinguishes a primary Anchor Adult from additional Support Adults. This creates consistency without making one person the young adult’s only source of support.

Compass Support Circle page showing one anchor adult, four support adults, circle guidelines, and a coordinator contact option
The full-circle view makes roles, backup support, and accountability understandable instead of implied.

04: Messages

Keep early communication connected to context

Messaging is tied to a specific request or activity and displays the relationship phase. Group conversations and a plain-language monitoring notice create more visibility while relationships are still developing.

Compass Messages page showing monitored group messaging for a group meal during phase one
Communication remains contextual, observable, and easy to report if a concern develops.

Evidence

This concept created questions to test, not answers to claim

The prototype is not evidence that the model is safe or effective. Its value is that it makes the assumptions visible enough for young adults, practitioners, and safeguarding experts to challenge them before a larger investment.

The next step would be to test whether people understand the relationship phases, feel they retain meaningful control, trust the coordinator’s role, and believe the safeguards would actually help rather than create friction or false reassurance.

What I learned

  • The obvious product pattern can be the wrong starting point for a sensitive system.
  • Safety and agency should be designed together rather than treated as opposing goals.
  • Good early prototypes can expose risky assumptions before they become expensive features.

What I would test next

  • Whether young adults prefer request-based support to profile-based matching.
  • Whether the support-circle model feels supportive or overly managed.
  • Whether staged communication is understandable and proportionate.
  • Which safeguards experienced practitioners believe are missing or unrealistic.