Every merchant onboarded is new, incremental payment and delivery volume, not volume taken from a competitor.
Even the 21% who do sell online overwhelmingly do it through their own website (85.8%) or a marketplace listing (45.3%), channels no AI agent can read today. The real addressable market isn't just the offline 79%; it's close to the full 32.3M, online or not.
Source: Eurostat Structural Business Statistics (2022); Eurostat/Statista SME e-commerce figures (2023)
BiziBees earns a share of the transactions it orchestrates. It is not a licensed payment institution or a courier operator.
A recurring fee once a merchant is actively listed across ACP, AP2, UCP and MCP surfaces: the ongoing cost of staying discoverable and query-ready for AI shopping agents.
A platform commission on gross merchandise value flowing through BiziBees, charged regardless of how the order was paid or delivered: the core marketplace-style take rate.
A markup on top of the underlying processor's rate (Adyen, Mollie or Stripe) for every card transaction routed through the platform, separate from the % of sales line.
A margin on each delivery booked through the local or global courier network, on top of the courier's own rate.
| Adyen Agentic | Stripe + OpenAI (ACP) | Square / Toast / Lightspeed | BiziBees.ai | |
|---|---|---|---|---|
| Target merchant | Enterprise, Adyen clients | Merchants already on Shopify/Etsy + Stripe | SMEs already on a modern POS | No-website brick-and-mortar SMEs |
| Onboarding | Enterprise sales-led | Self-serve, but only if already on Shopify/Etsy | Self-serve, needs a POS already | Self-serve, phone camera to start |
| Non-technical buyer channel | — | — | — | WhatsApp and voice, no AI literacy needed |
| Fulfillment | — | Merchant handles it themselves | — | Local + global courier network, orchestrated |
| Agent-protocol ready | Yes | Yes, ACP-native | Not yet | Yes, phased in as data matures |
Sources: Stripe Newsroom, "Stripe powers Instant Checkout in ChatGPT" (2026); rye.com agentic-commerce landscape roundup (2026)
Prove the core loop in one pilot market from day one across both no-system-required capture paths, photo capture and file import together, not one channel gating the launch. Real merchants, real orders, multiple ways in from the start.
Turn POS and e-commerce integrations into a formal partnership program rather than one-off engineering per platform. Payments and logistics partners are added the same way, market by market.
List agent-ready merchants across ACP, AP2, UCP and MCP as data quality matures, running in parallel with Phase 2's scaling, not stacked behind it.
Phase 2 and Phase 3 run in parallel, not end-to-end. Reaching AI shopping agents doesn't wait behind the partnership rollout.
A deceleration curve grounded in real merchant-network comparables (Toast, SumUp), not a straight-line extrapolation.
Year on year: 6.0x → 4.0x → 2.9x, decelerating rather than compounding. Year 4 revenue of roughly €4.24M is blended across subscription, commission, payments and logistics margin.
AI-native vertical SaaS and embedded-payments companies trade at a premium to generic SaaS, with Adyen at 8.1x revenue marking the top of the range.
Illustrative. Actual valuation depends on growth trajectory, margins and market conditions at the time. Sources: saasvaluationmultiple.com Fintech SaaS Multiples (2026); Windsor Drake Fintech Valuation Multiples (2026).
12 to 18 months of runway to prove the pilot-market loop and reach Phase 3.
Get in touch