Zamski shows you what you are buying before you wire the money, and whether the story still matches every quarter after. Technical due diligence from the target company's operating history, not the snapshot assembled in a data room.

What sits in a Zamski report?

Every sealed report states what the record confirmed, what it contradicted, and what we could not observe. Every claim resolves to a specific artifact in the target's own systems.

How is Zamski different from a data room review?

A data room is a snapshot assembled by the seller. Zamski reads what the company created without knowing anyone would one day examine it. Code, messages, tickets, calendars, AI sessions. Read together, they become a record no prompt can recreate.

Private equity and venture deal teams running technical diligence, and fund CFOs who route to them. Post-close monitoring on portfolio companies and pre-sale prep are the two adjacent motions.

R&D GRADE · Specimen page

Code-up software diligence, sealed weekly.

Every figure on every card resolves to a sealed git, ticket, CRM, or session record.

Three R&D GRADE report cards on synthetic companies. On a live engagement the pills open to the underlying record; here they are labels.

All companies on this page are synthetic demonstrations. No client data appears anywhere. How the grade is computed →

SPECIMEN 01 · MERIDEX · FAILING · DILIGENCE REVIEW · $34M EV · RUBRIC v0.5.0

ZAMSKI · R&D GRADESYNTHETIC DEMONSTRATION

FAILING

Meridex

AI contract intelligence · W1 · diligence review

Six engineers, 129 merged changes in a month, a demo that closes rooms. 16% of capacity touches the product customers pay for. An F that photographs like an A.

Growth allocationcapacity classification not connectedengineering leadership, via workstream headcount mappingREPORTED

Deal commitmentswhat revenue is waiting on9 blockers, 0 touchedREPORTED

Key-person riskwho walks out with the company3 subsystems, 1 authorSINGLE-POINT

Facts sealed 2026-07-06 · window 2026-06-08 to 07-05 · every figure resolves to a record. GIT fct_m1a204CRM fct_m1e05bGIT fct_m1f930
Negative space. ticketing and runtime telemetry not connected. The questions this grade cannot answer are stated on the card, with who can. An honest gap beats a confident guess.

This is an F that photographs like an A. Six engineers, 129 releases in a month, a demo that closes rooms, and 16% of the capacity touches the product customers pay for. The deck cannot show you the 16%. The evidence can. Every figure on the card resolves to a sealed fact.

SPECIMEN 02 · LANEWICK · W23 · DILIGENCE REVIEW · $58M EV · RUBRIC v0.4.0

ZAMSKI · R&D GRADESYNTHETIC DEMONSTRATION

DIVERGING

Lanewick

AI freight-audit platform · W23 · diligence review

Lanewick's numbers look right and look recent. 58% of capacity lands on revenue work; the twelve-week average is 31%, and the shift began five days after the data room was requested. The pitch says proprietary models; the spend shows one vendor and no training workloads. Question the window first.

Grade history

W18

W19

W20

W21

W22

W23

Growth allocationrevenue-linked work58% in window, 31% trailing 12-weekSTRONG

Maintenanceupkeep of the asset8% capacity, 0 security changesUNDER-INVESTED

Delivery riskcan they ship what they generate74% top-reviewer shareBOTTLENECKED

Key-person riskwho walks out with the company86% ingestion single-authorCONCENTRATED

Facts sealed 2026-07-06 · window 2026-06-08 to 07-05 · rubric v0.4.0 · every figure resolves to a record. GIT fct_6e02b1AI fct_39d7c4GIT fct_7c19e2
Negative space. CRM, runtime telemetry, and AI sessions not connected. The proprietary-model claim is management to evidence; the substrate reports the spend mix, not the training work.

Read the history strip. Lanewick sat at C for three issues, printed B at W21 (five days after the data room was requested), then fell back to C at W23. Same substrate, different behavior in different weeks. When a grade improves exactly when diligence begins, question the window first.

SPECIMEN 03 · OXBOW · HEALTHY B · PRE-SEED · $900K · RUBRIC v0.6.0

ZAMSKI · R&D GRADESYNTHETIC DEMONSTRATION

ALIGNED, WITH FINDINGS

Oxbow

AI voice-agent evaluation · W1 · pre-seed profile · diligence review

Two founders converting capital into product at a healthy clip. 74% of capacity lands on the declared September milestone. One claim does not match the graph: the deck says proprietary judge models; three of every four evaluation calls route to a rented one. Fix the claim or finish the model; either clears it.

Milestone velocityis capital becoming product74% capacity on declared milestoneSTRONG

Build realitydo the tech claims match the graph74% of eval calls to one external vendorPARTIAL

Exploration disciplinediscard is health at this stage33% AI generation discardedEXPLORING

Knowledge capturecan a future team inherit the heads38% decisions substrate-residentTHIN

Facts sealed 2026-07-06 · window 2026-06-08 to 07-05 · rubric v0.6.0 PS profile · every figure resolves to a record. GIT fct_ox1a02AI fct_ox1c44SESSIONS fct_ox1e19
Negative space. CRM, ticketing, and runtime are Not Applicable at pre-seed per profile and drop from the denominator; they join the registry as the company grows into them.

A healthy B under the pre-seed lens. Discard reads as exploration at this stage; founder concentration does not fire the floor rule. The reading catches one claim gap: the deck says proprietary judge models, and three of four evaluation calls route to a rented one. A specimen showing what the method reports as healthy.

Illustrative · not real companiesThree deal moments

Sealed reports. Every figure resolves to a record.

Three mock companies, three deal moments. Open one to read the full sealed report the deal team, the fund operating team, or the underwriter would receive. All numbers illustrative; none of these are real companies.

Evaluating someone else's organization? See the R&D GRADE ↑

Specimen book

Seven cards. One folder.

The full specimen book includes seven R&D GRADE cards across four synthetic companies: Meridex (F, W1 diligence), Lanewick at W20, W23, and W28 (the dressed-window and post-fix monitoring arc), Oxbow (B, pre-seed), and Tallgrass under the seed and Series A profiles as a stage-lens comparison. Every company in the book is synthetic. Enter your work email; we send the PDF within one business day.

Your email joins the Zamski journal subscriber list, tagged as a specimen-book request. You may receive a journal welcome email before the specimen book arrives. Unsubscribe from any journal issue in one click.

See this on your target. The pre-purchase pilot is the 30-Day Finding: read-only OAuth, one sealed report, deleted after 30 days or continued into a subscription.

Start a pre-close report

Running your own org? Start a 30-Day Finding.

Code-up software diligence, sealed weekly.

Sealed reports. Every figure resolves to a record.

What the deal team reads before the wire.

What the operating team reads before the quarterly with the fund.

What the underwriter reads before the next S-1 prep call.

Seven cards. One folder.

How Zamski uses cookies