• Sold to Visma in April 2025 — classic exit velocity when founders see GPT-5 on the horizon.
• Claiming '99% accuracy' on document separation when Claude can do it for $0.003 per page.
• Competing with Vic.ai and Stampli, who are all racing to become the same AI-native accounting middleware.
• Raised €6M Series A in 2023, right before every accounting firm got a ChatGPT Plus subscription.

12-18 months until OpenAI or Anthropic ships native invoice processing that integrates directly with DATEV and Xero. Visma bought defensive position, not offense.

• Company claims 'time savings of up to 70%' — the phrase 'up to' doing heroic work in that sentence.
• 'AI assistant Fin' — naming your AI feature after the thing sharks have when they're hunting you is... on brand.

If Finmatics was a human, it'd be the accountant who learned Python in 2023 and now tells everyone at dinner parties they're 'basically a data scientist.'

• Visma moat: 1,200 accounting offices already integrated is painful switching cost.
• DATEV ecosystem lock-in: German/Austrian tax compliance is genuinely complex regulatory territory.
• 120 employees means actual implementation knowledge, not just wrappers around OCR APIs.
• Serving accounting firms, not end-users — B2B2B insulation buys 12-24 months vs consumer AI tools.

The difference between Finmatics and ChatGPT Vision? ChatGPT Vision doesn't need a €6M Series A to read a receipt.

• Ship native integrations with GPT-4V and Claude 3.5 Sonnet APIs — become the orchestration layer, not the OCR vendor.
• Leverage Visma's enterprise distribution to bundle into existing ERP contracts before standalone AI tools fragment the market.
• Build proprietary training data from 1,200 accounting offices — the learned booking logic is defensible; the OCR is not.