Skriptr AS

Governing message. Skriptr clears Gate 1 market proof and scores a mandate-fit 17/30 at a debatable seed read, but it warrants AFTER DD — not a deal — because the entire seed framing and the ~NOK 90–100M ask (≈90–250x current ~NOK 4–25k MRR on an unbuilt LMS moat against a 6–12-month NotebookLM incumbent window) hinge on three document-checkable unknowns reachable today: the JUVO predecessor IP/user/revenue transfer (phase-determining STOP if unproven), payment-processor verification of the self-reported ~150 paying users, and the Antler convertible cap-stack and exact pre/post-money — resolve these and the deal either validates at seed or collapses to pre-seed.

Strongest reason to say no (IC must engage with it): "You are being asked to pay a ~NOK 90-100M forward valuation (10x an unachieved NOK 1M ARR, i.e. roughly 90-250x the current ~NOK 4-25k MRR) for a 6-month-old AS whose only register-verified moat-relevant facts are a VAT registration and a consultancy NACE code, while its single durable defensibility claim (LMS switching-cost moat) is unbuilt and its moat window is being closed by Google NotebookLM Norwegian support within 6-12 months."

Facts

Company key

NO-936713831

Sector

consumer_edtech

Register status

Registered (Enhetsregisteret, class 2)

Register facts carry class-2 provenance (the Nordic advantage). Every claim under this run resolves to its source via the scorecard cells below.

Scorecard — every cell clicks to its source

Anti-bias checks: deck-polish ✓ · local-preference ✓. The score is evidence-anchored judgment — a starting point for your debate, never an answer key.

Conflicts & the deciding DD questions

Conflict register — surfaced, never smoothed

• CONF-01 Operating-history conflict: founder narrative claims ~12-18 months of operating history and 150+ product iterations, but the register shows SKRIPTR AS was registered only 2025-12-06 — a legal entity ~6 months old at evalDate (2026-06-14). The self-reported operating maturity materially exceeds the register-verified entity age. (sources: register.registeredAt=2025-12-06 (class 2) vs register.legalForm=AS vs claim c-history-reconcile (class 3) vs claim c-team (class 3): 5 co-founders, ~150 iterations vs claim c-traction-users (class 3): ~150 paying / ~500 beta users) → Surface as the primary self-report vs register conflict, not smoothed. Most pre-AS history likely sits in a student-company predecessor (JUVO). Route as a deciding DD item: obtain JUVO incorporation/closure records, asset/IP transfer agreement and user/revenue continuity proof to confirm the operating history transfers cleanly into SKRIPTR AS. Until reconciled, treat the 12-18mo history as T3 and let it pull the phase read toward the pre-seed alternative.
• CONF-02 Classification conflict: register primary NACE code is 71.129 'other technical consultancy' (a services/advisory code), not a software-product code (62.01), while the company self-describes as an AI-native SaaS product. The register classification contradicts the self-described product identity. (sources: register.naceCode=71.129 (class 2) vs claim c-reg-nace (class 2) vs claim c-reg-formaal (class 2): vedtekter cite software/AI solutions vs claim c-positioning (class 3): SaaS workflow product vs ask.useOfFunds: 10x-ARR SaaS framing) → Flag as a register-lag conflict that weakens the SaaS-product narrative on the most decision-critical (Product/Market) dimensions. Note that the statutory purpose (vedtekter, class 2) does corroborate AI-EdTech software, partially offsetting. Cross-check against the revenue model in DD: confirm revenue is recurring product subscription (SaaS) vs consultancy/advisory billing, since the NACE choice may signal how the founders actually invoice today.
• CONF-03 Headcount conflict: founder claims 5 full-time-committed co-founders, but the register shows no registered employees (antallAnsatte absent). The self-reported committed full-time team is not corroborated by employee registration. (sources: register.employees=null (class 2) vs claim c-reg-noemp (class 2): no registered employees vs claim c-team (class 3): 5 co-founders, full-time committed) → Note as a common-at-pre-seed but unverified conflict: founders often work via convertible/sweat equity without registered employment, so absence of registered employees is consistent with a clean founder-run entity rather than disproof. Do not treat 'full-time committed' as verified. DD: confirm founder commitment via employment/founder agreements, vesting and any salary/Innovation Norway payroll records to substantiate the Team dimension.
• CONF-04 Date-detail conflict (minor): claim states incorporation 2025-12-02 with registration 2025-12-06, while the register digest exposes only registeredAt=2025-12-06. The stated incorporation date is not independently shown in the register fields provided. (sources: register.registeredAt=2025-12-06 (class 2) vs claim c-reg-incorp (class 3 narrative on date): incorporated 2025-12-02) → Low-materiality; both dates place the entity at ~6 months. Reconcile against the stiftelsesdokument date in DD if entity-age precision affects runway/clock calculations. Does not change the phase read.
• CONF-05 Moat-window conflict: founder originally claimed a 2-3 year defensibility/moat window, but external evidence (Google NotebookLM Norwegian support, UiO public case Apr 2026) compresses the effective moat to ~6-12 months. Self-reported durability contradicts the competitive-reality read. (sources: claim c-competition (class 3): NotebookLM Norwegian, UiO Apr 2026, ~6-12mo window vs claim c-positioning (class 3): workflow-depth wedge vs general LLM vs claim c-lms-moat (class 4): LMS-integration switching costs not yet built) → Surface as a self-report vs external-evidence conflict on Market/Product. Treat the analyst's compressed 6-12mo window as the working assumption and the founder's 2-3yr claim as optimistic T3. The compensating workflow-incumbency moat (Canvas/Inspera LMS) is class-4 and unbuilt — do not credit it. Feed into pre-mortem and the 'strongest reason to say no'.
• CONF-06 Valuation-figure inconsistency within the ask: structured valuation states NOK 90,000,000 while the useOfFunds narrative anchors '~NOK 100M post-money'. The deal's own headline number is internally inconsistent. (sources: ask.valuation.amount=90000000 (class 3, document-read) vs ask.useOfFunds text: '~NOK 100M post-money anchored on 10x-ARR' vs mandate.valuationCap=null) → Flag the 90M vs 100M discrepancy as an unresolved terms ambiguity. Per the vague-terms rule the valuation/instrument is anchored but the inconsistency plus convertible post-money 'TBD' (c-capital) keep Terms decision-critical: auto-promote exact pre/post-money, convertible cap/discount stack and resulting fully-diluted ownership to deciding DD. Both figures imply a 10x-ARR multiple on an unmet NOK 1M ARR target — a forward, not current, valuation basis to challenge in red-team.

The deciding DD questions

1. (decides traction) What is the bank-/accounts-verified current MRR, paying-customer count and monthly churn for SKRIPTR AS today, and does the claimed ~150 paying users (NOK 100-300/mo) represent recurring product subscription revenue rather than consultancy/advisory billing — given the register's NACE 71.129 'technical consultancy' classification and VAT registration as the only independent turnover signal?
2. (decides market) With Google NotebookLM already supporting Norwegian (UiO case Apr 2026) compressing the moat window to ~6-12 months, what is the demonstrated, unit-economically viable distribution channel that bridges the ~40-250x gap from ~NOK 4-25k current MRR to the NOK 1M MRR Nov-2026 Series A trigger — and what CAC/conversion evidence exists beyond the unallocated ~NOK 2M across 5 untested channels?
3. (decides team) Can the JUVO predecessor incorporation/closure records plus a signed asset/IP/user-base transfer agreement confirm that the claimed ~12-18 months of operating history, 150+ iterations and ~500 beta / ~150 paying users transfer cleanly into the ~6-month-old SKRIPTR AS — and do founder/employment agreements with vesting substantiate the 5 full-time co-founders given zero registered employees?
4. (decides terms) What are the exact pre- and post-money figures (resolving the NOK 90M vs ~NOK 100M discrepancy), the cap/discount/maturity terms of the ~NOK 2M Antler-affiliated pre-seed convertibles, and the resulting fully-diluted ownership and cap stack after the proposed NOK 10M equity round — and how is a 10x-ARR valuation defensible against an unachieved NOK 1M ARR target (~90-250x current annualised revenue)?

Pre-mortem — how this fails in 24 months

• Incumbent erosion: Google NotebookLM (and other general LLM tools) close the Norwegian-academic-writing workflow gap faster than Skriptr deepens its wedge, collapsing the ~6-12 month moat window and stalling paid conversion before the unbuilt LMS switching-cost moat ships. (Market)
  Falsifiable: Net new paying users decelerate or churn rises in any consecutive 2-month window after a major incumbent Norwegian-feature release; or NotebookLM/UiO-style institutional adoption is cited by lost prospects. · Early indicator: Month-on-month organic signup rate flattens or declines; rising 'why not just use NotebookLM' objections in beta feedback loop; CAC creeps up as organic word-of-mouth saturates.
• Distribution never clears the economics: the ~NOK 2M budget across 5 unallocated channels fails to deliver the ~125k signups (at 1% conversion) implied by the NOK 1M MRR target, because paid acquisition CAC exceeds NOK 100-300/mo LTV in a thin niche. (Traction)
  Falsifiable: By month 6-9 post-raise, no single channel demonstrates a repeatable, payback-positive CAC:LTV cohort at scale; MRR remains an order of magnitude below the NOK 1M Nov-2026 trigger. · Early indicator: Blended CAC reported above first-year revenue per user; channel tests rotated without one being scaled; MRR run-rate tracking <NOK 250k entering Q3 2026.
• Series A trigger missed: current ~NOK 4-25k MRR cannot bridge the ~40-250x gap to NOK 1M MRR within the runway, so the company exhausts the ~NOK 3M + this round without hitting the financing milestone and faces a down round or bridge. (Traction)
  Falsifiable: MRR is below ~NOK 400-500k by Q3 2026 (an off-trajectory read for a Nov-2026 NOK 1M target), with <6 months runway remaining and no Series A term sheet in hand. · Early indicator: Quarterly MRR growth rate insufficient to compound to NOK 1M by Nov 2026; cash runway dips under 9 months without a milestone-credible pipeline.
• The durable moat stays a hypothesis: Canvas/Brightspace/Inspera LMS integration (Q4 2026 roadmap) and multi-vendor LLM abstraction are never delivered or delivered late, so the product remains a thin wrapper exposed to single-vendor (OpenAI) pricing/policy shocks and trivial substitution. (Product)
  Falsifiable: By end-2026 no LMS integration is in production with at least one institution, and no multi-vendor abstraction layer exists; an OpenAI price/policy change materially hits unit economics. · Early indicator: LMS integration milestones slip or drop from roadmap; engineering capacity diverted to firefighting; gross margin pressure from foundation-model cost changes.
• Predecessor-history reconciliation fails: the ~12-18 month operating history and ~150-iteration / ~150-paying-user claims do not survive register/contract verification (JUVO student-company IP, user base or revenue did not cleanly transfer to the 6-month-old AS), invalidating the seed-stage read. (Team)
  Falsifiable: DD finds IP, user accounts, or revenue were not assigned to SKRIPTR AS, or the claimed paying-user/iteration counts are not corroborated by payment processor / VAT turnover data. · Early indicator: No clean IP-assignment or asset-transfer documentation from JUVO; VAT-reported turnover inconsistent with the ~150-user × NOK 100-300/mo claim; founder narrative dates conflict with register filings.
• Valuation/cap-stack overhang kills the next round: the ~NOK 90-100M post (10x an unachieved NOK 1M ARR, ~90-250x current revenue) plus undocumented Antler-affiliated convertible terms create a cap table that later investors will not clear without a reset. (Terms)
  Falsifiable: Convertible post-money/cap terms surface as conflicting or punitive on conversion, or a Series A investor requires a flat/down round versus the ~NOK 90-100M mark. · Early indicator: Post-money on existing convertibles remains undisclosed at DD; new-investor indications price below the current entry valuation; conversion mechanics dilute beyond founder/seed expectations.

IC red-team — attacks the deal (not self-validation, DEL 93)

Gate 2 recommendation

Condition / re-evaluation. AFTER DD — proceed to a focused, document-driven diligence sprint whose 2–4 questions decide everything, each reachable by data-room/register access today (evidence-access upgrade, not a future event): (1) Predecessor-continuity (CONF-01, phase-determining): produce JUVO stiftelsesdokument, IP-assignment / asset-transfer agreement and user+revenue continuity proof confirming the ~12–18mo operating history, ~150 iterations and user base assigned cleanly into SKRIPTR AS — if not produced, immediate STOP and downgrade to the pre-seed alternative. (2) Verified traction (c-traction-users vs c-mrr-target): payment-processor MRR export and active paying-subscriber count to resolve the internal tension (150×NOK100–300 ⇒ ~NOK15–45k MRR vs stated ~NOK4–25k); VAT proves >NOK50k turnover, NOT 150 active subscribers — treat the lower MRR band as working figure until resolved. (3) Terms (CONF-06): exact pre/post-money, Antler-affiliated convertible cap/discount stack and resulting fully-diluted ownership; reconcile the NOK90M-structured vs ~NOK100M-narrative discrepancy — Terms cannot clear without this. (4) Demand + distribution: the signed university intensjonsavtale document (or strike c-loi from Market), plus any current payback-positive channel cohort evidencing CAC < first-year ARPU. The forward NOK1M-MRR/distribution-economics question is a trajectory item parked on the watchlist clocks, not a gating access question.

Evidence Floor is met (all five conditions hold), so a Gate 2 recommendation issues validly. Gate 1 market proof is present — a live MVP iterated from beta feedback, VAT-registration as a register-verified (class 2) turnover signal, and a paying-user base — clearing the single hardest-working criterion despite most traction being class 3–4. The scorecard (raw 17/30; phase-critical Traction 2.5 / Product 3 / Market 2.5 under the seed lens) is a starting point for the investor's debate, not a verdict. The decisive call is the ON-HOLD boundary test: every deciding unknown that gates the proceed/stop decision is an evidence-ACCESS upgrade reachable today, not an evidence-EVENT that has not yet occurred. The phase-determining JUVO predecessor reconciliation is checkable now via IP-assignment/transfer docs and VAT turnover; the keystone 150-paying-user/MRR tension is resolvable via a payment-processor export; the cap-stack and 90M-vs-100M valuation inconsistency are resolvable via convertible term documents; the deprioritised university LOI is a document to produce or strike. Per the decision rule (deciding unknown reachable by access today → AFTER DD), this routes to AFTER DD, not ON HOLD — the forward NOK1M-MRR trajectory and distribution-economics proofs are genuine future events but they are watchlist-clock matters, not the immediate gating questions. The red-team's strongest reason to say no (a ~NOK90–100M forward valuation at ~90–250x current revenue on an unbuilt LMS moat against a 6–12mo incumbent window) is serious and must be carried into the DD as a price/terms challenge, but it does not justify a PASS ahead of the document checks that could either validate the seed read or collapse it to pre-seed; the STOP triggers (predecessor-transfer failure, unverifiable paying-user count) are wired into the focused DD so the deal kills itself fast if the claims do not survive verification.

Validation clocks — the assumption register

Each clock resolves at expiry to promote / partial / retire-with-propagation / extend — these become scheduled decisions in Tasks, not to-dos in an inbox.

Method & caveats

Source-class mix (share of material claims): class 1 0% · class 2 31% · class 3 44% · class 4 25% (n=16).

Evidence Floor [G-B]: met → Gate 2 validly issued.

DEL 93: no panel scores are used as evidence anywhere in this run; the source-class mix and the flags are the honesty mechanism.

Run a structured evaluation

Run the full evaluation funnel on this company — Gate 0 → six-dimension scorecard → pre-mortem → red-team → recommendation — against your fund's mandate. It runs in the background (a few minutes); the result lands here in the company file.