Welcome back to the EUVC Podcast, where we bring you the people and perspectives shaping European venture.
This week, Andreas Munk Holm is joined by Max Schertel, co-founder & CEO of finmid, and Tim Rehder, General Partner at Earlybird, to unpack the rise of embedded lending infrastructure for B2B platforms.
From food delivery and PSPs to ride-hailing and fleet platforms, finmid lets marketplaces offer financing directly to their merchants with a single integration, across 30+ European markets.
Together, they break down why embedded lending is often new capital, not just smoother UX; how better data lets you underwrite the “invisible” SME segment; and what it really takes to scale regulated infra across a fragmented Europe.
Here’s what’s covered:
01:03 What finmid does: One integration for platforms to offer any financing product to business users across Europe
02:02 Why embedded wins: Tim on data access, risk scoring, and turning platforms into “banks in all but the balance sheet”
04:05 Owning infra, not capital: Regulation, operations, and data engine vs outsourcing pure funding to institutions
06:43 Economics & margins: Market size, 60%+ gross margins, and why net income beats headline spread
10:47 Customer examples: How Wolt Cash works, proactive offers in the merchant dashboard, and +80% retention uplift
12:32 Impact on the market: New capital for underserved SMEs vs just smoothing the bank journey
17:57 Ticket sizes & duration: Typical loans of €10–20k, up to ~12 months, 85% renewal and the path to larger, longer credit
21:15 AI & risk: Using generative and agentic AI in ops (adverse media) and data science (millions of data points, daily model iteration)
29:20 Scaling to 30 countries: U27 + UK, CH, IS – regulation, payments rails and why “ugly detail work” is the real moat
40:17 Partner alignment: Making financing core to platform metrics (GMV & retention) and hard-won lessons on incentives
✍️ Show Notes
finmid in one glance
What they are:
Embedded lending infrastructure for B2B platforms – marketplaces, PSPs, fleet/ride-hailing, vertical SaaS.What they enable:
With one integration, a platform can offer working capital and other financing products to its business users across 30 European markets (EU27 + UK, Switzerland, Iceland).What they own:
Regulatory setup & licenses
Underwriting, data engine & risk modelling
Operations & workflows end-to-end
What they don’t own:
The balance sheet.
Capital comes from private debt funds, institutional investors, and (increasingly) banks who want yield, not distribution.
The goal: let a platform act “as if” it were a bank for its merchants while finmid handles all the messy infra underneath.
Embedded lending: from thesis to practice
Tim’s thesis on embedded B2B lending:
Platforms like food delivery apps, PSPs or ride-hailing marketplaces sit on rich, live transaction data for thousands of businesses.
An embedded lender can:
Plug directly into that data,
Underwrite risk more precisely, and
Distribute financing through the platform’s existing channels.
Instead of a restaurant being sent to a bank where:
Underwriting relies on static, historical financials, and
The bank has no view into real-time platform sales,
finmid sits inside the platform’s stack and:
Scores the merchant based on in-platform revenue + external signals (e.g. bank data via PSD2),
Pushes proactive offers via the merchant dashboard, email, or even physical mail / WhatsApp, and
Repays via revenue-linked, embedded collections.
It looks like the platform is the lender but the infra, risk engine and capital orchestration are finmid’s.
Economics & alignment: everyone wants the SME to grow
Max breaks down the model into a “triangular” relationship:
The platform
Wants healthy, high-revenue merchants.
Benefits from GMV growth and better retention when merchants can invest in inventory, marketing or expansion.
The merchant/fleet
Needs capital for short-term cashflow and long-term growth.
Values speed and convenience over squeezing every basis point from the rate.
finmid
Wants to support good businesses and build a high-quality loan book.
Optimises net income, not just margin – sometimes lower margin but higher conversion is the winning trade.
Today, finmid runs 60%+ gross margins, but Max is clear:
You don’t optimise for spread in isolation. You optimise for net income over time, driven by smart pricing and risk.
Because acquisition is effectively done via the platform, CAC is spent once on the platform and then amortised across thousands of merchants.
Who actually gets financed?
This is not just nicer UX on existing bank products. In many cases, it’s new access to credit:
Small restaurant chains and independent venues whose transactional data has never been used for underwriting.
Merchants with “thin files” in traditional banking, but rich in-platform and bank account data in reality.
Fleet owners and drivers who can now finance vehicles inside the ride-hailing or fleet management platform.
Banks typically:
Look at static, historical data (filed accounts, collateral),
Struggle with small ticket sizes, and
Have little incentive to build fine-grained models for a tiny restaurant vs a large corporate.
finmid’s approach:
Use platform transaction data as the primary signal.
Enrich with bank data (PSD2) and other digital traces where relevant.
Build dynamic risk models to serve smaller tickets, shorter terms and “non-obvious” borrowers.
The result: capital reaches SMEs who were either declined, ignored, or poorly served by banks.
Ticket sizes, terms & renewal
To keep risk controlled while models mature, finmid has set deliberate constraints:
Typical loan size today:
Range: €1k – €250k+
Core cluster: €10k – €20k
Typical duration:
Up to 12 months.
Duration is linked to finmid’s prediction of how long the business will stay active on the platform and its expected revenue.
Renewal behaviour:
~85% of businesses renew their financing once repaid.
That creates a recurring revenue pattern rather than one-off loans.
There is strong demand for longer terms and higher amounts – but Max is explicit:
In lending, you can earn a lot today and crash the portfolio tomorrow. We’re intentionally limiting ourselves to where the data is robust, and then extending from there as our forecasting improves.
AI in operations & risk: concrete use cases
Max outlines two big AI fronts: operations and risk modelling.
Operations — agentic workflows instead of manual lookups
Example: adverse media screening before payouts.
“Old world”: manual checks in a few databases, maybe LexisNexis, plus human judgment.
“New world”: agentic AI workflows that:
Crawl and interpret a far broader set of sources,
Return better results faster, and
Do it cheaper than any human-only process.
Risk & data science — millions of data points, daily iteration
Early days: hire a traditional bank risk expert to bring a “drawer of models”.
Reality:
Underwriting here is a data science problem, not a static rules problem.
Human creativity is a bottleneck on how many features and model variations you can explore.
With modern AI tooling, finmid can:
Continuously generate new model hypotheses,
Back-test them against historical performance, and
Deploy better versions quickly.
Combine that with the fact that finmid simply has more and better data than most banks in these segments, and you get an underwriting engine that:
Learns faster,
Iterates more, and
Can serve segments traditional players won’t touch.
Scaling to 30+ European markets: ugly work as a moat
finmid is now live in:
All EU27 countries, plus
UK, Switzerland and Iceland.
Why go so broad, so fast?
Large platforms (like Wolt and others) think pan-European from day one.
If Bulgaria or Latvia is important for them, they want lending there too, not just in France, Germany, or Spain.
What it actually took:
Regulation:
Understanding and navigating local rules, edge-case laws and legacy quirks (e.g. how direct debits work in Latvia vs Italy).
Payment rails & currencies:
Mapping the movement of money across 30 different banking and payments systems.
Languages & local behaviour:
Product and communication in local languages and currencies.
Understanding that in some countries physical mail works better, while in others WhatsApp or other channels outperform.
One telling example:
A single finmid team member trained 160 account executives at a partner platform in one go.
Those AEs now talk to merchants locally, in their language, with finmid’s product embedded in their day-to-day sales conversations.
Max calls a lot of this the “ugly work” such as reading regulations, debugging payment flows, adapting communication to channel quirks. But that’s precisely what becomes the defensibility:
“If you’re willing to do the ugly work properly, you can offer a simplicity that others can’t.”
Partner strategy: core metric alignment or bust
finmid’s distribution depends entirely on partners. So the key question is:
Do you move the platform’s core metrics?
For platforms, those are typically:
GMV (do merchants transact more?), and
Retention (do they stick with the platform longer?).
finmid has shown:
Financing users can have up to 80% better retention vs those without credit.
Merchants often grow their GMV once working capital constraints are eased.
But getting there wasn’t simple. Both Max and Tim stress:
Early on, you don’t have proof – just thesis and external references (e.g. similar products in the US).
You need champions inside the platform who believe in the model.
You must execute flawlessly on early pilots to generate hard data on GMV and retention uplift.
Tim shares a cautionary tale from another portfolio company:
Great product, indirect distribution via partners, but financial incentives weren’t aligned with the partner’s core business.
Result: the partner didn’t push the product the way it needed to be sold.
Lesson:
Being an “add-on revenue line” isn’t enough.
Your product must reinforce the platform’s core business, or you’ll always be second priority.
Founder–market fit: why Earlybird backed Max & finmid
From Tim’s perspective, two traits stand out:
Founder–market fit
Max and his co-founder Alex spent years at N26, seeing firsthand what it means to scale financial services across Europe.
They’ve lived the pain of regulation, localisation and infra complexity.
Persistence in a regulated, partner-dependent space
B2B lending infra is not a “weekend hack + go viral on Product Hunt” play.
Distribution takes long sales cycles and repeated “no’s” from large organisations.
Regulations and capital partnerships add another layer of difficulty.
Tim’s summary:
“You need founders who stay extremely focused and persistent. The journey is never easy, but Max and Alex have proven they can do both product and distribution in one of the toughest verticals.”
💡 One-Liner Takeaway
Embedded lending for B2B platforms isn’t just a UX upgrade — it’s a new credit rail for Europe’s SMEs, powered by better data, AI-native underwriting and partners whose core business grows when their merchants do.
