Illustrative use case · DTC beauty
Fund repeat purchase
A beauty brand with repeat-purchase customers can separate the economics of new acquisition from the working capital needed for inventory and fulfillment.
Cohort Capital is a human-led, AI-accelerated credit and growth platform. Our partners own every decision that moves capital. Our agent bench compresses cycle time, expands diligence coverage, and monitors the portfolio continuously — auditable end to end.
Operators and investors who have built, financed, and scaled consumer businesses — accountable for strategy, capital, credit, risk, and fiduciary duty.
Founder, Principal & Executive Chairman of Van Dyke Acquisitions. His work spans building, acquiring, operating, and financing consumer businesses.
15+ years leading marketing, e-commerce, and performance teams. Former global head of performance marketing and a recognized thought leader on AI-enabled growth and digital operating systems.
15 years leading operations, finance, compliance, and risk for leading performance-marketing agencies and growth organizations.
Experience financing and supporting consumer businesses, brought to a focused cohort-financing platform.
Capital-deployed and brands-backed figures reflect activity at the founding team's prior firm. These are historical activity figures, not Cohort Capital fund returns or a commitment to fund.
The agent bench does not replace Investment Committee judgment, legal review, or fiduciary responsibility. It compresses cycle time, expands coverage, raises monitoring frequency, and builds an auditable operating system across origination, underwriting, funding, servicing, reporting, and compliance.
Illustrative use case · DTC beauty
A beauty brand with repeat-purchase customers can separate the economics of new acquisition from the working capital needed for inventory and fulfillment.
Illustrative use case · Apparel & accessories
For a seasonal brand, a collection share tied to the financed cohort’s actual gross profit can flex with customer performance instead of a fixed daily debit.
Illustrative use case · Subscription commerce
A subscription brand can use customer-level retention, acquisition cost, and margin data to help determine how much incremental acquisition is financeable.
Illustrative use case · Agencies & platforms
An agency can introduce a brand with proven acquisition economics. Cohort handles underwriting and servicing under the agreed referral terms.
Illustrative use case · Home & lifestyle
A brand carrying several advances can request a review of its existing obligations alongside a proposed cohort-financing structure. Refinancing depends on underwriting and existing terms.
Illustrative use cases, not customer testimonials or historical results.
Each agent is a named function with a clear mandate, measurable targets, and — where capital, credit, or law is involved — a human gate. Together they form a single, learning operating system.
Scans the DTC universe for ICP fit on observable signals — commerce-stack detection, ad-library presence, hiring and review velocity, traffic growth — and produces a ranked pipeline before human outreach begins.
Manages agency, platform, and ecosystem referrers; attributes funded cohorts to sources and tracks each partner's contribution to funded volume and portfolio quality.
Personalizes outbound, screens basic eligibility — $250K+ monthly revenue, 6+ months trading, gross-profit, CAC and contribution-margin signals — and books the initial data connection.
Orchestrates read-only connections across Shopify, Meta, Google, Stripe, and Plaid; assembles the diligence data room and turns a multi-week process into an hours-based workflow.
Reconciles platform-reported revenue and gross profit against bank settlement and processor data; flags wash orders, related-party transactions, refund manipulation, and attribution gaming. The agent flags; a partner confirms.
◆ Human-gatedConstructs customer cohorts, computes CAC by channel and contribution margin after COGS, fulfillment, fees, and returns, and produces verified GP/CAC inputs for the underwriting engine.
Fits survival, hazard, sBG and BG/NBD models to forecast lifetime cohort gross profit with uncertainty bands — using the lower band to size advances and preserve downside protection.
Runs the credit scorecard, enforces the investment gates, and pushes verified inputs through the Cohort Waterfall and Fund Returns models to produce an indicative structure, confidence score, and written rationale for the Investment Committee.
◆ Human-gatedAssembles branded credit memos, underwriting exhibits, base/downside/severe cases, diligence summaries, and recommended terms into an IC-ready packet.
Designs and runs geo-holdout, matched-market, conversion-lift, ghost-ad and budget-split experiments, computing true incremental CAC versus blended or platform-attributed CAC.
Runs synthetic control, CausalImpact, difference-in-differences and media-mix modeling where experiments aren't feasible — and calibrates every model against real experiment results.
Models the rising marginal-CAC curve to identify the spend band that remains incremental, profitable, and financeable — keeping capital out of over-saturated media.
Recommends product, advance, share, cap, and duration-indexed multiple, and runs base/downside/severe cases — always inside the fund's defensible risk envelope.
Generates master agreements, per-cohort funding schedules, closing checklists, and executed-term records, and maintains the documentation system of record.
◆ Human-gatedManages borrowing-base reporting, covenant compliance, cash forecasting, and recycling between tiers — monitoring loss-adjusted asset yield versus debt cost and recommending leverage throttles as spreads compress.
Reconciles gross-profit and remittance flows; flags short payments, missed remittances, reconciliation exceptions, and cure requirements.
Re-pulls actual cohort performance weekly, refreshes remaining-collection forecasts and marks, and calculates realized-versus-underwritten variance across every cohort and the full book.
Monitors realization variance, CAC drift, retention compression, concentration, and payment delays — and recommends renewal pauses, repricing, reserve adjustments, deleveraging, or escalation.
Tracks exposure by brand, channel, vintage, platform, category and product, and computes expected loss, stress loss, concentration limits, and systematic platform-factor exposure.
Calculates live gross and net IRR, MOIC, DPI, RVPI, fee and carry accruals, and operator spread — reconciled to the Fund Returns model.
Continuously compares realized outcomes against original forecasts and recalibrates PD, LGD, retention, CAC and advance-sizing logic — turning underwriting accuracy from a static assumption into a compounding advantage.
Produces quarterly letters, capital-account materials, borrowing-base certificates, lender updates, and brand-consistent investor communications.
Answers LP, lender, and counterparty diligence using a governed knowledge base, source citations, and approved disclosure language.
Enforces pre-funding disclosure gates, monitors state commercial-financing laws, and flags terms drifting toward fixed-obligation patterns that require legal review.
◆ Human-gatedRuns the pipeline state machine, assigns agent tasks, enforces human approval gates, escalates low-confidence outputs, and keeps the operating cadence synchronized end to end.
Maintains the shared data substrate, feature definitions, model versions, lineage, and access controls — the operating moat that lets the platform learn across cohorts, brands, channels and vintages.
Tests agent outputs, enforces confidence thresholds, blocks unauthorized actions, and ensures no money moves and no credit or legal decision is finalized without the required human approval.
◆ Human-gatedAgent role titles describe functional mandates within a human-supervised system, not employment or legal authority.