In late 2025, we wound up Hokodo. After eight years, ten countries, more than €500 million in financed invoices and 100,000 business buyers served, and over €50 million equity raised, we made the decision to close.
This post is our honest account of why we built Hokodo, what we're proud of, what we got wrong, and what the experience taught us about where the real opportunity in B2B trade credit actually lies. That insight is what led us to found Liquidity Lab, and we think it's worth sharing.
Why we set up Hokodo
The three of us — Richard Thornton, Louis Carbonnier, and Sami Ben Hatit — came from inside the problem. Between us we’d spent years in financial services or software development, including at Allianz Trade where we saw thousands of companies struggling with their Accounts Receivable and unable to access finance and/or credit insurance. We knew the plumbing of B2B trade credit better than most.
We saw a stark gap. Around 60% of global B2B commerce runs on credit terms, yet the infrastructure for managing that credit was built for an offline world. Trade credit insurance penetration among SMEs stood at roughly 0.25%, leaving most B2B merchants highly exposed to non-payment risk. For a digital-first merchant trying to offer payment terms at checkout — instantly, automatically, at scale — the tools simply didn't exist.
Our thesis was that if you could bundle real-time credit decisioning, fraud detection, financing, collections, and insurance into a single API, you could give B2B merchants what consumer brands had taken for granted for years: instant, seamless payment terms at the point of purchase. We knew no-one had built a solution so far, and the need was clear. What could possibly go wrong?
What we're proud of
Hokodo achieved things we remain genuinely proud of.
We proved the technology worked. We built proprietary credit scoring, fraud detection, and a collections engine from scratch, and it worked. Our credit and fraud performance was excellent. We became the first B2B BNPL provider in Europe to hold a full EU payments licence and we processed payments for over 100,000 business buyers across ten markets.
We attracted serious people to the problem. Over seven years we raised over €50 million in equity — from Anthemis, Mosaic Ventures, Notion Capital, Korelya Capital, BNP Paribas's Opera Tech Ventures, and Citi — alongside a €100 million debt facility from Viola Credit. We built partnerships with SCOR, AIG, BNP Paribas, Citi, Munich Re, and others who don't take bets lightly.
We helped merchants grow in real, measurable ways. Ankorstore, one of Europe’s most successful B2B marketplaces, credited Hokodo with helping acquire over 200,000 loyal business customers. Across our merchant base, we consistently saw 40% higher conversion rates and 30% larger basket sizes when instant payment terms were offered at checkout. The product worked.
And we helped move a market. When we started, B2B BNPL barely existed as a category in Europe. By the time we closed, there were credible competitors across the continent, and we're glad to have been part of starting that.
What we got wrong
We owe an honest account of the mistakes too.
We took too long to find our ideal customer profile. We tried to serve everyone from small marketplaces to large enterprise merchants, and the resulting product complexity slowed us down. The companies that got the most value from Hokodo were mid-market digital merchants with high frequency repeat transactions — merchants in food services, auto parts, or travel. We should have known that earlier and gone deeper faster.
We were also too ready to believe we’d achieved product market fit (PMF). In the 2021/22 environment where new competitors were launching every day we mistook small wins for definitive signals of PMF because we were worried about losing the market to new entrants. As a result we made expensive hiring decisions before they were justified, and we paid a heavy cost for those mistakes.
The fundraising environment also shifted. We raised our Series B in June 2022, just as the rate cycle turned and VC appetite for fintech dried up sharply. In a different capital environment, the runway to Series C would have looked very different. Timing matters, and with hindsight, we should have cut costs dramatically and hoarded our Series B raise.
And our most ambitious pipeline deals — the enterprise-scale merchants that would have transformed the business — were months away from closing when we had to make hard decisions. The company needed revenue and runway simultaneously, and we couldn't keep both. While our credit and fraud performance were impressive, our financing cost was too high to bring us to profitability before 2027.
A fundamental insight: large enterprises don't want to outsource their AR
The market we were most excited about — large enterprise merchants — turned out to be structurally unlikely to outsource their accounts receivable to a fintech. And it makes complete sense, once you understand why.
Their cost of capital is far lower than ours could ever be. They don't need a third party to fund their receivables book, they can do it themselves, more cheaply, without giving up the economics. We tried to accommodate this by offering delayed payouts to reduce the cost of financing, but there were other issues.
Enterprise merchants have years of proprietary data on their customers: payment history, dispute patterns, relationship context, sector risk. No external provider can easily replicate that. Their credit decisions are often relationship decisions, which they aren’t comfortable outsourcing. They need automated credit decisions, yes, but they need to control them.
And their AR is deeply embedded in their ERP systems, their commercial strategy, and their customer relationships. Extracting it to a third-party creates integration complexity that is rarely worth the benefit.
So the real question was never whether large enterprises wanted better AR outcomes — they absolutely do. DSO is too high. Collections consume too much time. Credit approvals are too slow. More of their customers deserve payment terms. The question was whether they would get there by outsourcing, or by transforming from within. And the answer, increasingly, is the latter.
AI changes what's possible in-house
The arrival of capable, deployable AI fundamentally changes the calculus. The advantages that a specialist trade credit provider like Hokodo had — automated credit decisioning, fraud detection, agentic collections — can now increasingly be built inside the enterprise itself, at a fraction of the historical cost.
AI agents can handle routine collections follow-up with persistence and professionalism, escalating to humans only when judgment is genuinely needed. Credit models can process the same signals we used and return instant decisions at checkout. Order processing can be automated from email, WhatsApp, and EDI without manual intervention. The barriers that once made outsourcing attractive are falling and what enterprises are left with is a genuine opportunity to keep their proprietary data advantage and their customer relationships, while getting the operational improvements they've always needed.
The catch is that building these capabilities well is hard. Knowing which levers to pull, how to design the credit frameworks, how to get AI from pilot to production, that requires a combination of subject matter knowledge and tech know-how which isn’t easy to find.
Why we set up Liquidity Lab
Closing Hokodo was one of the most difficult things we’ve had to do. We grieve the company, the team, and what it could have become.
But what the experience left us with was clarity: there is a large and underserved market for the expertise we spent eight years accumulating, delivered differently. Not as a product that takes over a merchant's AR, but as a partner that helps them transform it themselves. That’s why we are starting Liquidity Lab.
Liquidity Lab is a specialist consulting firm focused on AI-driven financial operations for businesses that sell B2B. We work with CFOs and their teams — and with financial services providers in the credit and AR space — to deliver measurable improvements across the order-to-cash and procure-to-pay cycle.
For B2B sellers, we replace slow manual credit approval with automated decisioning; deploy AI-powered collection agents that free finance teams for the exceptions that actually need them; embed frictionless payment terms into digital buying journeys; build fraud detection that verifies buyers instantly and invisibly; automate invoice validation to catch errors before payment; and capture orders from every channel directly into ERP. These aren't theoretical outcomes: they reflect operational improvements we delivered at scale during the Hokodo years.
For financial services providers — credit insurers, trade finance platforms, BNPL operators, invoice finance businesses, banks building B2B credit products — the challenges are different but equally pressing: how do you underwrite faster, at lower cost, with better risk management, as margins compress and competition intensifies? We help FS clients design credit risk frameworks and underwriting algorithms, structure and optimise receivables financing facilities, build fraud detection systems purpose-built for B2B trade patterns, and use AI to automate collections across their portfolios without sacrificing the relationship quality that B2B credit requires. This is work we're well placed to do because we've been on the other side of those conversations. We know what institutional debt investors want to see, how a well-run credit portfolio behaves under stress, and where the operational risks tend to hide.
What sets us apart is that we combine deep financial expertise with the ability to actually build and deploy. Most consultancies can advise. Most technology vendors can implement. Very few can design the credit algorithm, train the collection agent, map the process flow, and work alongside a team until it runs in production. We can because we did it for eight years.
What comes next
The problems that motivated us to found Hokodo haven't gone away. B2B commerce is accelerating online. Payment terms remain the single biggest friction point in the buying journey. Finance teams are still spending enormous amounts of time on processes that AI can now handle.
What has changed is where the solution lives. We believe the next decade of improvement in B2B trade credit will happen inside enterprises and financial institutions, and that the teams who get there fastest will be the ones with the best combination of domain expertise and AI capability guiding them.
That's what we're now building at Liquidity Lab. We're excited about the work ahead, and we'd genuinely love to talk if any of this resonates with challenges you're facing.
