CEO presenter notes

Slide 1
The next right step for everyday parenting moments.
Assumption · Internal financing hypothesis pending the monthly cash ledger.
Say
We are building bounded decision support for recurring everyday moments—not an AI parenting expert and not a clinical service.
Proof
The frozen v1 scope and 18-month validation plan make the first investment question explicit.
Do not hide
€400k and the 18-month runway remain provisional until the monthly cash ledger reconciles.
Likely follow-up
Expect investors to ask which exact protocol enters the first programme and who owns it.
Slide 2
When a hard moment happens, parents need an appropriate next step now.
Assumption · Problem hypothesis; evidence pending.
Say
The difficult job is choosing what to try during a transition, boundary, or routine conflict—not finding another article.
Proof
The 100-parent programme is structured to measure this job directly.
Do not hide
We do not yet have denominated parent research proving frequency, urgency, or willingness to pay.
Likely follow-up
Expect a request for interview method, recruitment criteria, and evidence that the job repeats.
Slide 3
Content, a chatbot, and an appointment each solve only part of the job.
Assumption · Category framing; no superiority asserted.
Say
We do not dismiss the alternatives. Each is useful; Harmonia tests whether a constrained workflow can close the decision gap between information and professional care.
Proof
Official sources confirm personalised memberships and AI parenting products already exist; the appendix carries the dated evidence.
Do not hide
Differentiation and superior outcomes are not established.
Likely follow-up
Expect ‘why not ChatGPT?’—answer with protocols, boundaries, evaluation, and feedback rather than model branding.
Slide 4
One Croatian-speaking cohort. One recurring non-emergency job.
Assumption · Working ICP hypothesis; not validated demand.
Say
We start with one Croatian-speaking cohort and one recurring non-emergency challenge so value and safety can be measured.
Proof
This scope matches founder access and the frozen commercial-test design.
Do not hide
The ICP is a product hypothesis, not proven demand.
Likely follow-up
Expect questions about exclusions, qualification, and how the first challenge is chosen.
Slide 5
Context in. One governed action out. Outcome recorded.
Product intent · Target product behaviour; implementation state pending verification.
Say
The workflow captures minimum context, selects a governed path, offers one bounded step, and records what happened.
Proof
The architecture appendix separates retrieval, generation, deterministic protocols, refusal, and escalation.
Do not hide
No shipped workflow or UI is being represented here.
Likely follow-up
Expect questions about protocol rights, model choice, logging, deletion, and human review coverage.
Slide 6
Immediate guidance is limited to defined non-emergency situations.
Product intent · Required design boundary; clinical, legal, technical, and privacy evidence pending.
Say
Defined non-emergency guidance stops when medical, developmental, abuse, violence, self-harm, safeguarding, or emergency indicators appear.
Proof
Child-centred AI and data-protection guidance provide the design baseline; source links are in Q&A.
Do not hide
Clinical ownership, thresholds, jurisdiction-specific routes, red-team evidence, and incident operations are not yet complete.
Likely follow-up
Expect a diligence request for the escalation map and harmful-output test protocol.
Slide 7
Expand one governed loop into the family support system.
Product intent · Future scope; not seed-period delivery.
Say
The full ambition remains a governed family support system: profiles, living plans, progress, professional summaries, education, quality time, broader ages, and multiple children.
Proof
The supplied app specification and screening, technique, and UI sources describe these connected capability groups.
Do not hide
This is product intent, not simultaneous seed-period delivery. Expansion depends on the first loop passing value, safety, retention, rights, and operating gates.
Likely follow-up
Expect questions about sequencing, what evidence unlocks each module, and which capabilities should remain deterministic.
Slide 8
We have problem access. The seed must earn product proof.
Target · Workshops are unverified internal evidence—not app validation.
Say
The seed funds a 100-parent, 90-day consented programme: a four-week active protocol with eight-week follow-up.
Proof
Targets are 60% activation, 35% week-four retention of activated parents, 15% paid continuation by week six, and zero unhandled high-severity events.
Do not hide
These are internal targets, not benchmarks or achieved results. €8.99 and €11.99 are price assumptions.
Likely follow-up
Expect denominator, recruitment-quality, adverse-event, comparator, and channel-capacity questions.
Slide 9
The round buys evidence gates—not an eight-year forecast.
Target · No EBITDA year; release gates need accountable approval.
Say
The return narrative is a sequence of evidence gates: governance and prototype, pilot, closed alpha, paid beta, then repeat, launch, or stop.
Proof
The send-ahead deck assigns the sequence across the 18-month target runway.
Do not hide
Owners, detailed dates, and monthly spend remain diligence items.
Likely follow-up
Expect a request for kill criteria and what happens if only some gates pass.
Slide 10
Family access, psychology, and commercial execution—with ownership gaps made explicit.
Internal evidence · Commitment, biographies, technical ownership, and governance roles require verification.
Say
Sven brings business development, marketing, and programme execution. Tihana brings psychology and family-support experience.
Proof
The source deck and integrated content draft state those roles.
Do not hide
Commitment, decision rights, biographies, adviser consent, and engineering/security/privacy ownership remain unverified.
Likely follow-up
Expect ‘who builds it?’ and ‘who owns clinical and security decisions?’
Slide 11
€400k to fund the proof—not the fantasy.
Assumption · Amount and 18-month runway remain pending a founder-approved monthly cash ledger.
Say
The provisional ask is €400k to fund one cohort, one governed workflow, one acquisition loop, and an 18-month evidence decision.
Proof
The main and appendix slides use one amount and one runway assumption consistently.
Do not hide
No EBITDA date, terms, financing commitment, or automatic follow-on is claimed.
Likely follow-up
Invite review of the pilot protocol, architecture, data map, and reconciled monthly driver model.