About Scanovich
Aleksandr Mordvinov — applied AI for cross-border commerce.
Scanovich is a solo practice: explainable systems for customs workflows, marketplace unit economics, and settlement infrastructure. This page documents what is in production, how engagements are scoped, and who should reach out next.
In production
Three working products, honest status labels.
TN VED / HS classification bot
liveTelegram bot that returns explainable HS codes for product descriptions — with GRI trace, confidence, and source references. EAEU TN VED today. The hard part isn’t classification, it’s making the output defensible to a reviewer.
Supplier spreadsheet normalizer
liveTurns messy supplier Excel / CSV / PDF into a single clean schema: row-level translation from supplier language to EAEU-rule product names, fields ready for HS classification, and a flag for whether each item needs mandatory certification, voluntary certification, or none. The cert detection is the part importer ops teams asked for most — it saves hours per shipment.
YM Analyzer
live SaaSHosted at app.scanovich.ai. Yandex.Market sellers upload raw cabinet exports and get SKU-level P&L after fees, logistics, storage, and ad boosts. The engine is general — same math works for Amazon, Ozon, Wildberries once the fee models are added. That’s the next wedge.
Why cross-border commerce, why now
The stack is held together by spreadsheets and human memory.
Cross-border commerce runs on operational mess. Importers move goods by stapling together supplier spreadsheets, customs archives, classification memory, and broker favors. Marketplace sellers live between cabinet dashboards that don't tell the truth about margin and Excel sheets that forget which SKUs are actually making money.
Every layer of this stack is begging for explainable AI: not a chatbot, not a dashboard, not a “revolutionary platform” — just workflows that take the messy inputs these teams actually have, produce outputs they can defend to a reviewer, and stay out of the way.
The practice builds one slice at a time. The first slice is customs — highest pain and the most structured data. The second is marketplace P&L — the engine generalizes. The third wedge follows paying signal from the first cohort.
Short version
Background
- 2023
Corporate marketplace leadership (technology partner of Yandex Market) plus independent commerce and EAEU compliance consulting.
- 2024
Scanovich full-time — applied AI product engineering and solution architecture.
- 2025
YM Analyzer shipped and picked up its first real users.
- 2026
Customs stack live; AttestRWA settlement attestation (SEABW 2026); scoped engagements open across customs and marketplaces.
Engagements
Who to reach out, in order of priority.
No fundraising at this stage — deliberate. Before capital, the practice wants the right operators and design partners in the room.
Priority this quarter
Pilot partners
Import operations, customs brokers, or Yandex.Market sellers — scoped pilot on operational data under NDA: agreed workflow, success criteria, and a fixed review window.
The practice validates architecture against production constraints. Two to five concurrent pilots is the target for Q2.
For Ozon, Wildberries, and what comes after
Design partners
Ozon and Wildberries analyzers stay placeholders on purpose — the first versions are co-designed with two or three sellers, not guessed from a spec. Same logic for Amazon or Shopify sellers who want a real P&L tool instead of a cabinet dashboard they do not trust.
Operators who prefer shaping a product from zero over waiting for v5 should reach out.
The long game
Advisors
Operators who’ve built in customs-tech (Flexport, Altana, customs brokerage SaaS), marketplace profitability tools (Helium 10, sellerboard, Jungle Scout), or B2B AI for operational workflows. The ask is one 30-minute call a quarter, honest criticism, and a direct line when direction stalls. Equity grant when the entity exists, if it’s mutual.
In return: structured access to product direction and shipping cadence, and direct input while the wedge is still being defined.
When the signal is real
Investors
Angels who write early checks into pre-revenue solo founders in the cross-border commerce space. The practice is pre-incorporation today; the entity will live wherever it makes sense for the first cohort of paying customers.
The first check will be small — runway enough to hire one engineer and keep velocity through the next twelve months. Conviction at month six beats FOMO at month one. If you fund at pre-revenue and like boring, structural SaaS plays dressed as AI, reach out now.
Public record
The alternative is writing the same email twenty times, and the right people tend to show up only when the signal is clear enough for them to recognize themselves in it. If that's you, use the contact section below — Telegram is usually same day, email within a business day.
Contact
Talk to us
Questions about pilots, partnerships, or deployment? We usually reply within a business day.
Direct
JST (UTC+9). Telegram — same day. Email — within 24h.