The consumer experience: intake, thoughtful introductions, curated events and human support. We understand the person, consider compatibility and find a comfortable setting to meet.
Commercial model: matchmaking, memberships, introductions and experiences.
02 / Lumiora
How the system learns.
The intelligence layer being built to structure readiness, compatibility and potential friction, with explanations a human can review.
Commercial direction: per-assessment fees, APIs and B2B licensing. A signed LOI provides an early pricing signal.
What we want to understand
Three questions. A more complete picture.
The individual
Are you ready?
Intent, self-awareness, emotional regulation, relationship patterns and how someone approaches conflict and repair.
The pair
Who fits beside you?
Values, faith, communication, emotional needs, complementarity and places where expectations may differ.
The life
Can your futures fit?
Marriage, children, money, lifestyle, career, responsibilities and health priorities.
The intended output is an explanation.
A structured view of readiness, areas of alignment, questions to explore and possible friction—supporting a decision rather than reducing a person to a percentage.
Voice in context
A signal to ask. Not a verdict.
The question tells us what to look for. The voice tells us where to look closer.
“We’re using it to see where the interview deserves a second question.”
Illustrative interview
“I communicate well.”
A follow-up about conflict brings a longer pause, a faster pace or a less specific answer.
The next step: ask for an example, understand the context and compare it with the rest of the conversation.
A voice shift alone does not establish dishonesty, personality, readiness or relationship risk. This is not lie detection or diagnosis. Whether any signal predicts a useful outcome remains a validation question.
For technical partners
A system designed to be examined.
The proposed pipeline connects interview context to human review, then to outcomes over time.
This is the intended architecture and development direction, not documentation for a production API.
Consent & structured intake
Record purpose, permissions, relationship goals and interview context before processing sensitive inputs.
Language & contextual voice features
Associate transcripts with question context. Explore pace, pitch, pauses, intensity and changes from a person’s own baseline.
Reviewable assessment
Organise observations, evidence and uncertainty. Human reviewers probe gaps, label findings and decide what warrants further discussion.
Compatibility & introductions
Compare values, needs and life expectations. Explain areas of fit and possible friction to support considered introductions.
Consented outcome feedback
Track what happened, what friction emerged and what helped. Use that feedback to evaluate and improve future assessments.
The research & data thesis
The moat is the longitudinal dataset.
Structured interviews, expert labels, real-world interactions and follow-up outcomes become valuable together—if the measurements hold up.
MeasureMatchObserveSupportMeasure outcomesImprove
Can we measure it?
Establish reliability, clear constructs and consistent human labels.
Does it matter?
Test whether a signal adds useful information beyond the interview and ordinary context.
Does it predict?
Evaluate against later outcomes, across relevant populations, languages and settings.
Early commercial signals · Reported in the founder’s pitch
What has happened. What remains to prove.
Signed LOI
Up to $3
Per automated assessment
An external matchmaking company has expressed willingness to pilot and pay for Lumiora. Final scope and commercial terms remain to be agreed.
Observed event
50 → 10
Attendees to sign-ups
Ten sign-ups from an event of fifty people: an observed 20% conversion in one event, not a forecast for future events.
Internal target
<$1
Automated analysis cost
A development target per assessment. Not yet proven, and not a current operating margin.
The pitch also reports organic demand, referrals and a first deposit. These are early signals, not product-market fit or evidence of predictive validity. The LOI is not booked revenue.
The next build
Rigour is part of the product.
Measurement, engineering & governance
The next capabilities to assemble span relationship science and psychometrics, AI and data science, product engineering, clinical referral expertise, privacy and partnerships.
Technical design priorities include explicit consent, restricted access, clear retention and deletion rules, traceable model versions, reviewable evidence and evaluation for bias.
These are requirements for the build—not claims of completed implementation, certification or regulatory approval.
Interested in the measurement challenge, the engineering roadmap or the business behind it? Start a conversation with Deena Al Jassasi, founder of Maktoub.