The Meridian — Assumption Test Cards

Osterwalder / Strategizer Format | Milestone 3 Deliverable

7 testable assumptions mapped to The Meridian's user flow: 3 core path assumptions (Discernment, Clinical Routing, Coaching) and 4 drop-off assumptions that represent the "missing tree" — states where the system fails to retain or route the user. Framework reference: False Positive Framework, Lead Domino Theory, Relational Mirror AI logic.

Assumption 1: The Discernment Path Medium Risk
Step 1: Hypothesis
We believe that
faith-curious professionals who complete the full discovery and track experience will reach a state of Integrated Discernment — the ability to distinguish clinical needs from spiritual needs and act on both.
Step 2: Test
To verify that, we will
run 8–12 users through a prototype discovery + track sequence (even paper-based) and conduct exit interviews measuring whether they can articulate the difference between a clinical issue and a spiritual one in their own life.
Step 3: Metric
And measure
% of users who can correctly self-triage a scenario (clinical vs. spiritual vs. both) in a post-track assessment, compared to their pre-intake baseline.
Step 4: Criteria
We are right if
≥60% of completing users show measurable improvement in discernment accuracy, AND they describe the experience as "clarifying" rather than "prescriptive."
Assumption 2: The Clinical Routing Path Medium Risk
Step 1: Hypothesis
We believe that
the AI Relational Mirror can reliably detect clinical red-line signals (addiction severity, self-harm indicators, clinical depression markers) during the discovery process and route those users to human clinical professionals.
Step 2: Test
To verify that, we will
create 20 scripted user scenarios (10 clinical, 10 non-clinical) and have Lauren + one additional clinician independently score each. Run the same scenarios through the AI logic and compare routing decisions.
Step 3: Metric
And measure
sensitivity (true positive rate for clinical detection) and false negative rate (clinical cases the AI missed). False negatives are the critical failure mode.
Step 4: Criteria
We are right if
AI achieves ≥90% sensitivity on clinical detection (we can tolerate false positives — over-referring is safer than under-referring) with 0% false negatives on severe cases.
Assumption 3: The Coaching Path Medium Risk
Step 1: Hypothesis
We believe that
users who don't respond to AI-delivered content alone will respond to a human coach, and that the AI's diagnostic data will make coaching sessions more effective than cold-start coaching.
Step 2: Test
To verify that, we will
identify 5 users from the prototype who stalled on content tracks. Pair them with a coach who receives the AI's diagnostic summary vs. a coach who starts from scratch. Compare session outcomes.
Step 3: Metric
And measure
sessions-to-breakthrough (how many sessions before the user reports a meaningful insight) and user-reported satisfaction with the coaching experience.
Step 4: Criteria
We are right if
AI-informed coaches reach breakthrough 40% faster than cold-start coaches, AND users rate the experience ≥4/5 on relevance ("they understood where I was").
— Drop-Off Assumptions (The Missing Tree) —
Drop-Off 1: "Not Ready" High Risk
Step 1: Hypothesis
We believe that
the Lead Domino sequence (Achievement → Image → Control → Autonomy) is calibrated correctly — users will tolerate the Relational Mirror probing past surface-level topics without bouncing.
Step 2: Test
To verify that, we will
run discovery sessions with 10 users and track the exact moment they disengage (change topic, give short answers, express discomfort). Map disengagement to the specific Lead Domino layer they were on.
Step 3: Metric
And measure
drop-off layer (which of the 4 layers triggers disengagement), time-to-disengage, and whether a "cool-down" prompt can re-engage them.
Step 4: Criteria
We are right if
≥70% of users progress past Layer 2 (Image & Perception) without disengaging, AND cool-down prompts recover ≥50% of those who stall.
Drop-Off 2: "Wrong Fit" High Risk
Step 1: Hypothesis
We believe that
the 4 False Positives (Achievement Mirage, Control Fortress, Image Mirror, Autonomy Island) are comprehensive enough to classify ≥80% of the target persona.
Step 2: Test
To verify that, we will
interview 15–20 "High-Functioning Void" professionals and have them self-identify which False Positive resonates. Track anyone who says "none of these."
Step 3: Metric
And measure
classification coverage (% who map to at least one FP), time-to-identification, and whether "none" respondents reveal a 5th pattern.
Step 4: Criteria
We are right if
≥80% self-identify with a False Positive within the first discovery session. If <80%, a 5th category likely needs to be defined.
Drop-Off 3: "Spiritual Allergic Reaction" High Risk
Step 1: Hypothesis
We believe that
the "implied faith, not forced" positioning allows us to introduce scripture and spiritual content without triggering the Religious Allergy in faith-curious (but not faith-committed) users.
Step 2: Test
To verify that, we will
create 3 versions of the same track content: (A) no faith language, (B) implied/indirect spiritual framing, (C) explicit scripture. Test each with 5 faith-curious users and measure reaction.
Step 3: Metric
And measure
completion rate per version, sentiment shift (positive → negative), and the specific content element that triggers disengagement.
Step 4: Criteria
We are right if
Version B (implied faith) retains ≥70% of users through to track completion, AND sentiment stays neutral-to-positive. If Version B also triggers drop-off, the spiritual integration timing needs redesign.
Drop-Off 4: "Silent Plateau" High Risk
Step 1: Hypothesis
We believe that
users who get relief from surface-level content (Layer 1: Achievement) will naturally want to go deeper, rather than plateauing at "good enough."
Step 2: Test
To verify that, we will
track 10 users through a multi-week content track. After they report initial relief or insight, observe whether they voluntarily continue or need prompting to go deeper.
Step 3: Metric
And measure
voluntary continuation rate (% who advance without prompting), days-to-plateau, and whether a "what's still unresolved" nudge re-engages plateau users.
Step 4: Criteria
We are right if
≥50% voluntarily continue past Layer 1 content. If <50%, the product may need a "bridge mechanism" between surface relief and deeper work — the initial value alone won't pull them through.