§ 08 ── ABOUT

We exist because methodology lives in partner heads.

Specialist funds compound when their methodology survives scale. That methodology — the scoring frameworks, founder-coaching patterns, evidence-base rigor, LP voice calibrations — is what distinguishes specialists from generalists. It lives in partner heads. We encode it as operational substrate.

§ 08.1 ── WHY

Specialist funds compound. Generalists average.

The funds that outperform over decade-plus horizons share a feature: they understand one segment, one stage, one geography, or one thesis deeply enough that their methodology is meaningfully different from anyone else's. The methodology is the moat. Generalists have capital and brand. Specialists have judgment.

But methodology that lives in two or three partner heads is fragile. It doesn't survive co-founder departures, 2× AUM growth, generational handoffs, or Fund II raises. Fund AI OS exists because methodology preservation is an operational problem, not a documentation problem. Documents don't run on autopilot at 2am. Substrate does.

FIG. 01 ── METHODOLOGY · LOCATION MATTERS · PARTNER HEAD vs SUBSTRATE

§ 08.2 ── WHY NOW

Three conditions converged.

FIG. 02 ── FRONTIER MODELS · AUDIT PRIMITIVES · SPECIALIST ECONOMICS → 2026

Frontier models are good enough

The current generation of frontier models is consistently strong on agentic workflows specialist funds need — multi-step diligence, evidence synthesis, communication drafting under voice constraints. They weren't five years ago. They are now.

Audit chain primitives matured

Hash-chained append-only audit trails are mainstream infrastructure now. EU regulators are starting to expect them implicitly. Article 9 SFDR funds need them explicitly. The substrate to make AI agentic systems regulator-defensible is no longer speculative.

Specialist funds need the lever

Sub-€500M AUM specialist funds have fundamentally non-scalable operations — the partner is the product. They can't grow into BCG-style large teams. Their only scaling lever is operational substrate. The economics finally work.

§ 08.3 ── PRINCIPLES

What we commit to.

Ten principles. The first eight have governed every engagement from day one. The last two — I15 substrate-mirroring and pattern migration — are the architecture's recursive commitments to itself and to the cohort.

METHODOLOGY PRESERVATION

Your methodology is yours (L2). We encode it. We don't sell it. No engagement makes your overlay part of our product without explicit pattern-migration consent.

IP CLARITY

L0 universal doctrine · L1 framework (operator IP) · L2 instance (yours). You own the production code, audit chain, customer overlay, eval suite, runbooks, and all data. Spelled out in the MSA.

AUDIT BY DEFAULT

Every action writes to the audit chain via the Audit Curator (sole-emission per tier). There is no path in our architecture where decisioning bypasses the audit primitive. Even partner overrides are audit entries.

CONSENT IS NON-NEGOTIABLE

Founder data flows are opt-in, granular, and revocable. No engagement weakens consent gateways. Revocation halts flows within 24 hours; aggregates purge within 72 hours.

HUMAN AT THE LIMIT

L8 architectural floors on Investment Principal (Tier-2) and Founder Support (Tier-2) are architectural, not policy. The autonomous path doesn't exist in the code. Human judgment stays in the loop where consequences are highest.

REVERSIBILITY BIAS

Originals archived before destructive operations. Append before rewrite. ADRs immutable. Ledger append-only. Amendments logged with rationale.

ANTI-RITUAL DISCIPLINE

The substrate serves the fund, not the reverse. Ceremony is proportional to stakes. Hygiene flags procedures that aren't earning their keep.

HONEST EXPERTISE

The substrate tells you what it knows, what it verified, and where it's likely to be wrong. Evidence modes labeled. Confidence bands explicit. No hidden assumptions.

SUBSTRATE- MIRRORING

NEW · I15 ── Every commitment we enforce on customers applies recursively to our own operations. We operate under the same audit chain, the same L8 floors, the same κ ≥ 0.85 verification, the same D7 §12 inspection. The substrate has no outside.

PATTERN MIGRATION

NEW ── Patterns that prove cross-cutting across multiple customers in the same L1 can migrate L2 → L1 — with explicit customer concurrence, attribution, and compensation. The substrate compounds across the cohort.

§ 08.4 ── CONTINUITY

The substrate travels.

Across partner transitions. Across model upgrades. Across vendor migrations. Across fund cycles. Across customers — patterns migrate L2 → L1 with attribution and compensation. The customer overlay is YAML, versioned in git, owned by you. The audit chain is yours. The eval suite is yours. The runbooks are yours.

FIG. 03 ── SUBSTRATE RIBBON · UNBROKEN ACROSS FOUR TRANSITION GATES

PARTNER TRANSITIONS

Methodology survives the co-founder departure.

MODEL UPGRADES

Overlay survives every LLM replacement. Schema replay preserves past decisions.

VENDOR MIGRATIONS

Substrate runs on Anthropic, OpenAI, whatever next. Provider-agnostic by architecture.

FUND CYCLES

Customer overlay transports to Fund II intact.

Anti-vendor-lock by architecture, not by promise.

§ 08.5 ── ANTI-FRAGILITY

The substrate compounds. AI tools decay.

Most AI tools depreciate. The prompts that worked stop working. The fine-tunes go stale. The dataset drifts. Vendor lock-in turns into vendor exit. Eighteen months in, the marginal value is lower than it was at deployment.

The substrate compounds. Every audit chain entry deepens the institutional record. Every replay test sharpens schema migration safety. Every κ ≥ 0.85 measurement refines calibration. Every Tier promotion validates methodology against observed outcomes. Every L2 → L1 pattern migration improves the framework for the whole cohort.

FIG. 04 ── SUBSTRATE COMPOUNDS · AI TOOLS DECAY · 5-YEAR HORIZON

Every replay improves calibration. Every decision deepens the substrate. Every cohort engagement strengthens the L1.

§ 08.6 ── FIT

Not every fund is a fit.

FIG. 05 ── FIT WINDOW · AUM SWEET SPOT · PARTNER COUNT 2-5

WE WORK WITH

— Specialist VC funds €20M – €500M AUM
— 2-5 partner teams
— Thesis-driven (impact / climate / healthtech / deep tech / etc.)
— Article 9 SFDR posture or equivalent rigor demand
— 7+ year fund life (substrate compounds)

WE DON'T WORK WITH

— Generalist funds (methodology doesn't compound for them)
— Funds below €20M AUM (we'd be too expensive)
— Funds above €1B AUM (they'll build in-house)
— Funds seeking copilot tools (different product class)
— Funds seeking enterprise rollouts (different scale)

§ 08.7 ── TEAM

Small. Senior. Specialist.

Fund AI OS operates as a senior-specialist team. We don't scale by adding junior consultants. We scale by encoding methodology — ours and yours — into reusable substrate (L1 frameworks) so each engagement gets faster, deeper, and more defensible than the last. The team will name itself publicly when the second engagement signs. Until then, ask us directly.

The model changes.
The substrate remembers.

── FUND AI OS · 2026

§ END ── ENGAGE

If this resonates with how you think about your fund — talk to us.

Read the methodology →Engagement model →Talk to us →