Skip to content

Alfine Labs

Business Plan

Alfine Labs · April 2026 · Confidential · Work in progress


1. Product Summary

Who: Serious musicians — amateurs, students, and professionals working toward mastery.

Problem: Slow, uneven, and uncertain progress toward goals.

How: Four AI-powered products that together cover the full system of how musicians improve.

  • Outer Ear — feedback on what happens in your practice session.
  • While Away — productive work for time away from the instrument.
  • Small Muscles — body conditioning built for the demands of music.
  • Ricercare — an adaptive skills development engine grounded in evidence of what works across Alfine's user base.

2. Founder

Product visionary with deep background in music, technology, and business.

Practitioner.

  • Five-year sabbatical to attend the Naples Conservatory; thesis recital; serious continuing amateur.
  • The user. Every product decision comes from having lived the problem from the inside.

Strategist.

  • Management consulting background; broad commercial and operational strategy terrain.
  • Comfortable with rigour and ambiguity; brings strategic and analytic discipline to the company.

Builder.

  • Day job: senior software engineer at a Series C startup building complex AI-driven event planning software.
  • Solo (AI-assisted) execution at the prototype level; team forms once funded.

Personal and professional network spanning the US, Italy, and India — music, business, and technology, including the MIT alumni network.

3. Portfolio Thesis

For centuries, musicians have assumed that the journey toward mastery is slow, uneven, and uncertain.

The five pains

The Perceptual Bubble. You practise inside a bubble of inaccurate self-perception — your ears mislead you and solitude flatters you.

The Flow Trap. The practice session that feels most like music is often the one that produces the least improvement — flow, the state every musician aspires to in performance, is the enemy of productive practice.

Practice Under Siege. Practice time is under constant threat — from your day job, your family, your environment, and your own psychology — and there's always a "good" reason not to practise.

The Recalcitrant Body. Your body is a shared resource between music and everything else in your life — it ages, accumulates damage, arrives at the instrument compromised, and may not even tell you what's wrong.

Conventional Wisdom. Your direction comes from traditions transmitted across generations, methods that have settled into authority, views passed through the sources you follow. Each is honestly given; none is grounded in evidence of what actually produces improvement.

Pain to gain

Each pain has its inverse:

The Perceptual Bubble → outside eyes and ears. You hear and see what actually happened, not what you thought was happening.

The Flow Trap → practice that resists flow. The counter-intuitive behaviours that actually produce improvement — interrupting, isolating, staying with difficulty — made habitual enough to deploy when flow pulls the other way.

Practice Under Siege → improvement away from the instrument. The time life gives you becomes time that moves you forward.

The Recalcitrant Body → a body that cooperates. Conditioning built for the demands music makes on you.

Conventional Wisdom → direction grounded in evidence. A developmental path built on what actually produces improvement in musicians.

How musicians improve: the underlying system

Musical improvement — or lack thereof — is produced by a system. The five pains are the five places that system fails.

Two of them live within the session. Perception (PP1) and practice behaviour (PP2) must both work for the session to produce improvement.

Three live outside it. Mental readiness (PP3) is what the musician brings from time away from the instrument. Physical readiness (PP4) is the body's capacity to execute what the session demands. Path-efficiency (PP5) is how well-optimised the musician's work is over months and years — what to practise, in what sequence, toward what goals.

Within the system, gains from outside the session amplify or mute gains within it. Gains in perception (PP1) and practice behaviour (PP2) within the session are amplified by having a cooperative body (PP4), a mind prepared away from the instrument (PP3), and an optimised path to follow (PP5).

How musicians improve: the underlying systemDiagram showing musical improvement as a system. The session sits at the visual centre as a flywheel of perception (PP1) and practice behaviour (PP2) cycling clockwise between each other. Path-efficiency (PP5) sits upstream as a vertical band on the left. Mental readiness (PP3) above and physical readiness (PP4) below modulate the session. All three modulate the session with bar-ending modulation lines. The flywheel produces improvement. Path- efficiency PP5 Mental readiness PP3 Physical readiness PP4 The session Perception PP1 Practice behaviour PP2 Improvement toward goals


4. Why Now

The pain points are old. The drivers of change are recent.

  • Push: AI is finally good enough. Multimodal AI (audio, video, score together) has matured to where credible practice feedback is now feasible to build. No product has done this yet for serious musicians — that's the opening.
  • Pull: sports has shown the pattern. In a generation, deep apprenticeship traditions in tennis, golf, and baseball have absorbed data and AI-augmented training without losing their character. Music is the near-perfect structural parallel.

5. Market Sizing

What's in the addressable market.

Global TAM: ~24M serious Western classical musicians.

Anchored in a US national survey showing 2.2% of US adults practise or perform classical music. Applied at 2% to mainstream-classical countries, 1% to countries with strong competing traditions.

Segment Share of TAM
Serious amateurs ~87–91%
Serious child learners (including teenagers) ~3–5%
Music teachers (private, K-12, faculty) ~3–4%
Pre-professionals (conservatory students) ~1.5%
Freelance performers ~1–2%
Salaried orchestra and institutional ~0.3–0.5%
Elite (international concert and recording careers) <0.1%

The 10:1 amateur-to-professional ratio is normal for serious adult avocations.

What's not.

  • Other genres. Jazz (~7M) and Indian classical (~2.5M) extend the multi-genre figure to ~34M.
  • Lapsed musicians. ~6–13M in the US alone — re-activation is a different motion.
  • Geographies. China, Middle East, and most of Africa.

6. Competitive Landscape

Two axes of competition.

  • Product overlap — companies whose products overlap with Outer Ear, While Away, or Small Muscles.
  • Audience capture — entities commanding serious musicians' time, money, and attention regardless of product overlap.

Rating benchmarks for the audience-capture table: - Reach — share of the 24M TAM ever encountering the entity. ● 50%+, ◕ 25–50%, ◑ 10–25%, ◔ 5–10%, ○ <5%. - Time — hours per week per median engaged user. ● 20+, ◕ 5–20, ◑ 2–5, ◔ 30 min–2 hrs, ○ <30 min. - Money — annual spend per engaged user, benchmarked against weekly-lessons cost (~€3–5K/year). ● matches or exceeds, scaling down to ○ negligible.

6.1 Portfolio-level 2×2

Portfolio competitive landscape A 2x2 chart plotting competitor categories on two axes. X-axis: product overlap with the Alfine portfolio. Y-axis: audience capture among serious musicians. Labels are generic categories in the plural; named entities appear in the tables below. High audience · Low overlap High audience · High overlap Low audience · Low overlap Low audience · High overlap no current competitor in both territories product overlap → audience capture → low high low high General-purpose AI assistants Serious-musician YouTubers Instrument manufacturers (latent) Online learning platforms Body-awareness training for musicians Music-education thought leaders Performing arts medicine practitioners Instrument-specific body-technique methods Beginner-keyboard AI coaches Audio-only practice-feedback apps Multimodal practice research projects Ear-training apps Theory apps Musician-specific body-management apps Musician-specific yoga apps Generalist MSK digital therapeutics Y-axis measures attention captured, not music-specific capture. General-purpose AI assistants are high not because musicians use them for music, but because musicians use them daily for everything else. Dashed dots: audience already captured (for non-musical or hardware purposes) plus resources to build a musician-specific product. Right-half x-position reflects portfolio-weighted fractional coverage. Top-right is empty — where Alfine Labs is building toward. Labels are generic categories; named entities appear in the tables below.

6.2 Product-category competitors

Harvey ball scale: ○ empty, ◔ quarter, ◑ half, ◕ three-quarter, ● full.

Outer Ear territory

Competitor Product Capability Audience Compatibility Business strategy Product strategy Technology strategy
ROLI Hardware + subscription, consumer beginner AI coach on simplified keyboard IR hand tracking + proprietary instrument
SkyNote / TELMI Academic research, no commercial entity Audio app shipped, multimodal in research Multimodal MIR, conservatory-validated
Audio-only practice-feedback apps (Violy, Clefer, Leopold AI, LLaQo) Freemium/subscription, cross-instrument Pitch and timing feedback Audio MIR, single modality

While Away territory

Competitor Product Capability Audience Compatibility Business strategy Product strategy Technology strategy
EarMaster (+ ear-training cluster) Subscription + B2B to music schools Structured ear-training drills, ABRSM-aligned Adaptive drill engine, multi-language
Theory apps (Teoria, musictheory.net, Hookpad) Freemium/subscription, cross-instrument Theory drills and reference Drill engine, written exercises

Small Muscles territory

Competitor Product Capability Audience Compatibility Business strategy Product strategy Technology strategy
Kaia Health B2B via employers and health plans Clinical MSK digital therapeutics Computer vision exercise correction, RCT-validated
Hinge Health B2B, scaled public company Dedicated PT + health coach per user Platform + human coaching at scale
Intermission Free iOS app + paid retreats Instrument-specific video library Static video, no personalisation
SM app/course cluster (corpSonore, Musicians Maintenance, Music Strong) Founder-led small businesses Musician-specific content, mixed formats Video, PDF, course platforms

6.3 Audience-capture competitors

Competitor Reach Time Money Workflow embedding Credibility
General-purpose AI assistants (ChatGPT, Claude, Gemini)
Serious-musician YouTubers (aggregate)
Instrument manufacturers with digital ecosystems (Yamaha, Roland, Kawai, Steinway, Casio)
Online learning platforms (tonebase, ArtistWorks, MasterClass music, Josh Wright Piano TV)
Body-awareness training for musicians (Alexander, Feldenkrais, Body Mapping)
Music-education thought leaders (Kageyama, Gebrian, Mortensen, and others)
Performing arts medicine practitioners
Instrument-specific body-technique methods (Timani)

6.4 Strategic synthesis

  • The portfolio is lightly contested overall. Outer Ear has adjacent competition (ROLI, in the beginner segment) and latent competition (instrument manufacturers, generalist AI providers).
  • Latent competitors are a real but localised threat. Bundling moves by instrument manufacturers cost Alfine addressable market in those segments only — the instrument market is fragmented enough across makers and instrument families that the damage stays bounded.
  • The practice session matters more to Alfine than to its threats. It's the entry point and primary data source for Alfine's portfolio thesis; for instrument manufacturers, it's downstream of their core business — they have the easier structural path, Alfine has the higher motivation.

7. Product Architecture

Four products, mapped to the system of musical improvement described in Section 3.

Pairwise technology overlaps. The three vertical products share a pairwise technology axis each:

  • Outer Ear ↔ While Away — score rendering and audio-to-score alignment.
  • Outer Ear ↔ Small Muscles — computer vision (technique analysis in Outer Ear, exercise form feedback in Small Muscles).
  • While Away ↔ Small Muscles — LLM daily-life reasoning (context-aware sessions in While Away, daily check-ins in Small Muscles).

Ricercare — the platform layer.

  • Ricercare draws on data from the three verticals to produce evidence-based direction across the long arc of a musician's development.
  • Insights flow back into each vertical, sharpening their feedback over time.
  • Ricercare ships in Stage 3, after the verticals have generated enough real-world data to train its model.

Coverage check. The four products cover the five pain points described in Section 3 — the within-session engine (PP1, PP2), the readiness amplifiers around it (PP3, PP4), and path-efficiency over the long arc (PP5).

Pain point Outer Ear While Away Small Muscles Ricercare
Engine
PP1 — Perceptual Bubble
PP2 — Flow Trap
Readiness amplifiers
PP3 — Practice Under Siege
PP4 — Recalcitrant Body
Path-efficiency
PP5 — Conventional Wisdom

8. Technical Challenges

The challenge: multimodal AI for music. Combining different signals — sound, motion, anatomy, score — into useful, accurate feedback. Several technical sub-problems within it:

  • Computer vision and video analysis. Tracking body posture, hand position, technique, and gesture from consumer-camera footage.
  • Gesture-to-sound fusion. Connecting what the body is doing to what the sound becomes — the harder integration problem, requiring different modalities at the right time scales.
  • Biomechanics integration. Adding domain knowledge about musician anatomy, instrument-specific load patterns, and injury-relevant body mechanics into the analysis.

How we approach it. Two minority positions:

  • Macroscopic feedback is enough. Feedback at the level of the first few hard-hitting observations a teacher makes in a lesson covers most of what serious musicians need. The majority view chases ever more nuance — analysing sound, gesture, and the causal relationship between them — interesting for elite musicians and academic research, but overkill for most musicians.
  • Human-provided criteria beat purely learned ones. Pedagogical expertise (from teachers, performance traditions, advisors) provides better evaluation criteria than end-to-end learned models, especially while the data substrate is small. Alfine builds with explicit human-derived criteria as load-bearing infrastructure, not as a bootstrap to be replaced.

9. Evolution Path

Alfine Labs reaches its full shape across three stages: first product, full portfolio, platform layer. The path runs through Outer Ear at stage 1, While Away and Small Muscles at stage 2, and Ricercare at stage 3.

Stage 1: First product — Outer Ear.

  • Outer Ear ships as a multimodal practice intelligence product: hardware mount, video and audio capture, frictionless self-review with limited AI assistance.
  • Software is founder-led with AI-assisted engineering.
  • Critical technical assumptions de-risked: video feasibility, audio capture quality, mount design.
  • End state: Outer Ear has paying customers; first recurring revenue.

Stage 2: Full portfolio.

  • While Away and Small Muscles ship as software-only subscription products.
  • Outer Ear evolves into multimodal AI feedback at the macroscopic level.
  • Shared infrastructure starts paying off — score across Outer Ear and While Away, computer vision across Outer Ear and Small Muscles.
  • End state: three products in market, cross-sell across one musician base, shared technology, and a real-world dataset on which to train Ricercare's adaptive learning model for musical skills.

Stage 3: Platform layer — Ricercare.

  • Ricercare ships as the platform-layer product, drawing on the cross-portfolio dataset accumulated in stages 1–2.
  • Its adaptive learning model for musical skills produces guidance grounded in what actually works — across the long arc of a musician's development.
  • Platform-layer insights flow back into each vertical, sharpening their feedback over time.
  • End state: Alfine is a platform company — Ricercare's adaptive learning platform can power institutional training, learning marketplaces, and assessment and certification.

The alternative: While Away first.

Path A: Outer Ear first Path B: While Away first
The bet Pursue first-mover advantage on highest-leverage product Build confidence and momentum on less-contested products
Stage 1 product Outer Ear While Away
Stage 1 capital Higher (hardware build) Lower (software-only)
Stage 1 technical risk Higher (multimodal AI, hardware) Lower (audio + LLM + score)
Stage 1 market size Largest of the three verticals Smaller — secondary pain category
Competitive context Contested market with big upside and narrow window for emerging pure players Less contested market with smaller upside; no current window pressure, but conditions could change
Expansion logic Outward from the session — natural Inward toward the session — harder
Defensibility once shipping Strong (workflow, data accumulation) Lower (more easily replicated)

Path A is chosen. Path B becomes right under specific conditions: severe capital constraints, or generalist competitors entering Outer Ear's territory before Alfine can ship. Neither is the current situation.


10. Revenue Model

Three revenue-model variants across four products.

  • Software-only subscription — While Away, Small Muscles, Ricercare.
  • Hardware-plus-subscription — Outer Ear.
  • Institutional licensing — per-seat, activated at scale.

Per-product launch pricing.

Product Model Standard tier Premium tier
Outer Ear Hardware + subscription TBD, below ROLI's ~£558 + £13/mo User-controlled criteria
While Away Subscription ~€14/mo ~€25–28/mo
Small Muscles Subscription ~€28/mo (prevention, management) ~€45/mo (adds performance optimisation)

All three: free trial, with a free tier reverting to non-personalised content if not converted.

Portfolio bundling activates from Stage 2. Discount against the sum of individual subscriptions; specific level set once retention data is in hand. Functions as cross-sell and retention mechanism.

Blended ARPU at Stage 2–3: ~€25–40/month, depending on single-product/bundled mix. Gross margin healthy on software; hardware thin until production volume.

Ricercare monetisation is open. Three candidates: premium tier across the portfolio, separate subscription, or B2B licensing to conservatories. Deferred to Stage 3.


11. Distribution and Go-to-Market

Early concentration.

  • Wedge candidates: musicians who play piano or harpsichord as a second instrument — organists, conductors, composers, early-music students with secondary-keyboard requirements. Acute, underserved pain.
  • Commercial segment: serious adult amateurs in classical music.
  • Geography: US, UK, Germany; Italy as founder-network base.

How Alfine reaches customers.

Channel
Outer Ear crowdfunding (hardware risk, innovation-tolerant audience)
Small Muscles performing-arts-medicine practitioners as referral channel
While Away content and community (unfamiliar category, audiences in adjacent content)
Cross-portfolio advisor amplification, teacher channel, community and content authority

How acquisition stays affordable.

  • Cold acquisition is hard once. Outer Ear acquires serious amateurs cold; subsequent products are sold to a known, qualified, engaged customer.
  • Cross-sell CAC declines with each portfolio addition. By stage 2 it's the dominant share of new sign-ups.

How distribution evolves.

  • Stage 1: cold acquisition through Outer Ear's launch channel.
  • Stage 2: cross-sell within installed base; portfolio-level channels strengthen.
  • Stage 3+: B2B emerges — teachers who already recommend Outer Ear become licensed users; institutional buyers come from individuals who first bought as musicians.

12. Team

Today. Solo founder, AI-assisted on engineering. Not actively recruiting a co-founder; open to a strong match if one appears.

By stage.

  • Stage 1 (Product). Solo founder; AI-assisted engineering; specialist contractors for hardware and any work beyond AI-assisted general engineering; advisor relationships with domain experts.
  • Stage 2 (Portfolio). Core leadership team. First hires own areas (engineering, product/operations) at leadership level; most execution remains contracted, outsourced, or AI-assisted.
  • Stage 3 (Platform). Full operating org.

13. Partnerships and Advisor Needs

Category Activates Role
Domain advisors Pre-launch Individual experts whose research or practice informs product (Molly Gebrian, Gary Karpinski, Meinard Müller for Outer Ear and While Away; Bronwen Ackermann and performing arts medicine practitioners for Small Muscles). Advisory relationships; specific targets and outreach timing tracked in the cofounder/advisor search document.
Authority-lending partners Stage 1 onward, per product Platforms and prominent voices whose endorsement reaches serious musicians with credibility no paid marketing can match. Varies by product: tonebase and prominent piano pedagogy voices for Outer Ear; Noa Kageyama (Bulletproof Musician) and musicianship educators for While Away; performing arts medicine associations and prominent physiotherapist-to-musician networks for Small Muscles. Engaged when each product earns the endorsement; not earlier.
Non-traditional institutional B2B Stage 3+ Freelance teachers, online pedagogy platforms, coaches, independent content businesses — the non-traditional music education economy. Later-stage B2B layer. Alfine Labs sells authority, pedagogical infrastructure, and verified-progress data to players who cannot produce them on their own.

14. Customer Discovery

Customer voice so far. The pain-point diagnostic and product architecture were grounded in plural sources: community forums, pedagogy literature, founder experience as a serious musician, and observation of other musicians in their practice contexts.

Formal discovery to date.

  • One interview (jazz musician).
  • Signal: behaviour-mirror meta-feedback was less compelling than direct AI feedback on playing. Held as live hypothesis pending more interviews.

Forward plan.

  • 15–20 interviews during the incubator, portfolio-wide.
  • Pain-point level: are PP1–PP5 live and ranked as framed? Does the practice-session-as-high-ground thesis survive how musicians rank their own obstacles?
  • Product level: for each product, does the solution resonate? What language do musicians use for the pain?
  • Findings feed into product gates, pricing, and the strategic thesis.
  • Discovery continues at lower cadence post-incubator, expanding into the eventual tester cohort.

15. Resource Requirements and The Ask

Pre-seed (Stage 1 — Product).

Ask. €300–400K from angel investors with relevant domain experience. Crowdfund supplements or substitutes.

Use of funds. - Outer Ear MVP and crowdfunding launch (~€90–190K). - While Away and Small Muscles MVP groundwork (~€50–80K). - Founder runway and operations (~€110–170K).

Milestones at end of stage. - Outer Ear shipping with paying customers; technical assumptions de-risked. - While Away and Small Muscles ready for stage 2 launch. - Customer discovery completed; pricing and retention signals in hand.

Seed (Stage 2 — Portfolio).

  • Funds While Away and Small Muscles launch, Outer Ear evolution to multimodal AI feedback, and the core leadership team.
  • Delivers three products in market with cross-sell, shared technology paying off, and the dataset for Ricercare's adaptive learning model.

Series A (Stage 3 — Platform).

  • Funds Ricercare build and launch, full operating org, and B2B expansion to the non-traditional music education economy.
  • Delivers Ricercare in market, portfolio at platform scale, and an institutional licensing pipeline.

16. Key Risks

Exogenous forces that could undermine the portfolio regardless of execution quality.

  • Foundation model providers expand into application spaces. Anthropic, OpenAI, Apple, or Google ship musician-focused capabilities at platform or OS scale. Alfine Labs' differentiation narrows.
  • Foundation model supplier-relationship shift. API pricing, access restrictions, commercial terms, or regional availability change in ways that disrupt product economics. All four products depend on this infrastructure.
  • Single-founder dependency during critical window. Founder incapacitation between W1 and W22 stalls the portfolio; material risk continues through crowdfund fulfilment at W44. No co-founder in place during this window.

17. Closing

Alfine Labs

Thank you.

Contact: [founder name] — [email] — [website]


Alfine Labs Business Plan · April 2026 · Confidential