"Innovative" is the most abused word in startup copy. So here is the version with sources: six specific claims, each with the evidence that supports it and the condition under which it would stop being true.
We implement the Agentic Commerce Protocol and publish an MCP catalogue, so an AI agent can complete a purchase end to end.[1]
Falsified if: a rival ships ACP plus MCP plus policy controls before we launch publicly in 2027.
Invoices are generated in FA(3) and submitted to KSeF as part of the order lifecycle, with offline24 fallback and UPO storage.[2]
Falsified if: KSeF is replaced or substantially deregulated before 2028.
WCAG 2.2 AA is enforced by CI against a legal obligation for Polish e-commerce that has been live since 28 June 2025.[3]
Falsified if: the accessibility obligation is repealed or exempted for e-commerce.
BLIK, Paczkomat, COD and invoice-and-pay-later are default-on because that is how the country actually pays: 1.4 billion online BLIK payments in 2025, 30 000 lockers.[4][5]
Falsified if: card and courier preference in Poland reverses durably.
LCP ≤ 2.5 s, INP ≤ 200 ms and CLS ≤ 0.1 are pipeline checks, not marketing numbers.[6] INP has been a Core Web Vital since March 2024 and most incumbent storefronts have not been retuned for it.
Falsified if: Google changes the metric set in a way that makes the guarantee meaningless.
Full fee schedule, real price-and-stock API, no dark-pattern upsells, Omnibus-correct lowest-price-in-30-days display.
Falsified if: we ship a fee we don't publish, or a discount badge that can't be traced to a real prior price.
Each row carries its source and a verification status. Verified means we checked it against the cited publication. Standard means it is a legal or technical instrument. Projection means it is somebody's forecast and is labelled as such.
| Metric | Value | Status | Source and note |
|---|---|---|---|
| Polish online retail turnover, 2025 | 91.971 bn PLN | Verified | [7] Krajowa Izba Gospodarcza (KIG), on GUS retail reporting Nominal growth of 6.8% year on year; online share of total retail rose to 9.1% from 8.9%. |
| Share of retail happening online | 9.1 % | Verified | [7] Krajowa Izba Gospodarcza (KIG), on GUS retail reporting Still roughly 6–9 percentage points below the Western-European average, which is the headroom a new platform is competing for. |
| Active online shops in Poland | ≈75 thousand | Verified | [8] Dun & Bradstreet Poland, via Rzeczpospolita 7.8 thousand opened and 5.4 thousand closed during 2025: plus ~200 net per month. |
| BLIK transactions, 2025 | 2.9 billion | Verified | [4] Polski Standard Płatności (PSP) Worth 441.5 bn PLN; 10 billion transactions since launch. Daily mean of 8 million, peaking near 13 million. |
| BLIK payments online, 2025 | 1.4 billion / 219 bn PLN | Verified | [4] Polski Standard Płatności (PSP) Online is BLIK's largest channel, up 27% in value year on year. Estimated card spend online: ~65 bn PLN. |
| Average BLIK online basket | 158 PLN | Verified | [4] Polski Standard Płatności (PSP) |
| Active BLIK accounts at end-2025 | 20.7 million | Verified | [4] Polski Standard Płatności (PSP) Each active user averages 12 BLIK transactions a month. |
| Growth of BLIK Pay Later (deferred payments) | ≈440 % y/y | Verified | [4] Polski Standard Płatności (PSP) Offered by four banks; 235.5 thousand transactions worth 59.4 mn PLN. |
| InPost parcel lockers in Poland | 30 000 machines | Verified | [5] Grupa InPost 70 000 lockers and 100 000 out-of-home points across Europe; ~1.4 billion parcels handled in 2025. |
| KSeF e-invoicing, phase 1 | 1 Feb 2026 effective date | Verified | [9] Ministerstwo Finansów / Krajowa Administracja Skarbowa Issuing required for taxpayers whose 2024 sales exceeded 200 mn PLN. Receiving invoices via KSeF became mandatory for everyone. |
| KSeF e-invoicing, phase 2 | 1 Apr 2026 effective date | Verified | [9] Ministerstwo Finansów / Krajowa Administracja Skarbowa All remaining entrepreneurs must issue structured invoices in KSeF, unless monthly invoiced sales stay at or below 10 000 PLN. |
| KSeF e-invoicing, final phase | 1 Jan 2027 effective date | Verified | [2] Ministerstwo Finansów Transitional reliefs end: the 10 000 PLN monthly exemption, cash-register invoices and 450 PLN receipts-as-invoices all close. Penalty provisions were scheduled from 1 Jan 2027; the ministry announced a deferral draft on 16 Sep 2026. |
| Firms already issuing in KSeF voluntarily | 306 000+ businesses (Mar 2026) | Verified | [10] Krajowa Administracja Skarbowa / gov.pl Before the April 2026 mandate took effect — a signal of how ready the market already is. |
| Mandatory e-invoice schema | FA(3) from 1 Feb 2026 | Standard | [2] Ministerstwo Finansów KSeF 2.0 adds offline24 mode, invoice attachments and KSeF certificates. Invoices are archived in the system for 10 years. |
| Polish Accessibility Act in force | 28 Jun 2025 effective date | Standard | [3] Dziennik Ustaw RP Implements Directive (EU) 2019/882. Existing service contracts may run unchanged until 28 June 2030 at the latest. |
| Regulator for e-commerce accessibility | Minister for Digitalisation | Standard | [11] gov.pl — Dostępność cyfrowa (Ministerstwo Cyfryzacji) Article 18 imposes specific duties on online shops: identification, security and payment functions must be perceivable, operable, understandable and robust. |
| Maximum penalty for non-compliance | 10 % of turnover | Standard | [3] Dziennik Ustaw RP Also capped at ten times the average monthly wage. A consumer complaint unanswered within 30 days is deemed granted. |
| Who is exempt | Micro-enterprises offering services | Standard | [12] Ministerstwo Funduszy i Polityki Regionalnej A dependency trap for merchants: a shop that outgrows micro-enterprise status inherits the obligation overnight. |
| Agentic Commerce Protocol released | 29 Sep 2025 open source, Apache-2.0 | Standard | [1] OpenAI Co-developed by OpenAI and Stripe, with Meta now listed as co-maintainer. ChatGPT Instant Checkout is the first implementation. There is a Stripe Shared Payment Token API for non-Stripe processors. |
| Weekly ChatGPT users | 700 million | Verified | [1] OpenAI The addressable audience for an agent-reachable catalogue — and an interface, not a marketplace. |
| Protocol transport | REST + MCP | Standard | [13] Agentic Commerce Protocol Checkout configuration can be published as a conventional REST API or as a Model Context Protocol server. Payment tokens stay out of the agent's reach. |
| Core Web Vitals pass thresholds | 2.5s / 200ms / 0.1 LCP / INP / CLS | Standard | [6] Google / web.dev Measured at the 75th percentile of real users. INP replaced First Input Delay in March 2024; most legacy storefronts have never been re-tuned for it. |
| Marketplace commission vs own-store payment fee | 8–20% vs 0–2% | Projection | [14] Industry compilation citing PwC Strategy& (secondary source) A widely repeated industry range, not a measured statistic. SwiftShop therefore publishes its own fee schedule rather than asserting a comparison. |
| Polish firms selling cross-border online | 4.9 % | Projection | [14] Industry compilation citing PwC Strategy& (secondary source) Cited as ~112 100 entities. The untapped export channel that a multilingual, multi-currency platform is built to open. |
Any founder can list reasons they are right. Here is the argument that would make us wrong, stated in its best form.
Poland's market is already served. Allegro dominates spontaneous brand recall, Temu is second by traffic, and around 75 000 shops fight for a market that grew only 6.8% in 2025.[8][7] Growth is converging on general retail growth, which means the land-grab phase is over.
Agentic commerce may stay small. ACP launched with single-item purchases in one country. Multi-item carts, broader regions and non-Stripe processors were still rolling out. It is entirely possible that agent-initiated checkout stays a rounding error for years, in which case our differentiating layer is a cost centre rather than a moat.
Compliance is not a moat, it is a floor. KSeF connectors are now sold by every Polish accounting vendor. Accessibility tooling is commoditised. If regulation is a race to the bottom, the winner is the cheapest integrator, not the most principled one.
Our answer. We agree with the first and third points and we don't think they help us — they help whoever executes fastest. The second point is the real risk, and it is why the platform is built as seven separable layers: if agents stay marginal, the same storefront still wins on accessibility compliance, KSeF-native billing and a checkout tuned for BLIK and lockers. If agents take off, the agent gateway is already there. We are not betting the company on one thesis being right.
You will not find a customer count, a logo wall or a testimonial on this site. SwiftShop.pl is pre-launch: there are no customers to name, so naming any would be a fabrication. The same rule applies to revenue, funding and headcount.
The full register is machine-readable at /data/market-data.json. If a number on this site can't be traced there, treat it as our opinion.
Each claim above names the condition under which it fails. If you think one has already failed, we would rather hear it now than at launch.
Every figure above is drawn from the sources listed here. The machine-readable register lives at /data/market-data.json.
SwiftShop.pl is pre-launch. Figures come from the sources below; projections are labelled as such.