Tech and circular economy
Dipli: a finance team that bet on AI to keep pace with growth
Dipli makes second life simple for tech products, from trade-in and refurbishment all the way to resale. In 2025, its finance team, led by Head of Finance Enzo Makaïa, chose to entrust the matching of all its payment flows to AI agents. With a team of just two, it keeps the books up to date as the business grows.
-
96%
of payments matched automatically from the first month
-
100+
orders automatically allocated from a single partner payout
-
Many
sales channels: online sales, retailers, telecom operators, B2B and trade-in programmes
-
2
people on Dipli's finance team
Contributors
-
Enzo Makaïa
Head of Finance, Dipli
-
Tanguy Pennel
Co-founder, Dipli
-
Amalia Bercot
Co-founder & CEO, Sanso
About the client
- Activity
- All-in-one platform for the tech circular economy: trade-in, returns, refurbishment and omnichannel distribution of electronic devices across Europe
- Founded
- 2016
- Headcount
- Around 45 people
- Finance team
- 2 people
- Flows handled by Sanso
- Online sales, retailers, telecom operators, B2B, trade-in programmes
- With Sanso since
- March 2025
- Connected tools
- PennylaneStripeQontoPartnersIn-house ERP
The challenge: varied sales channels, complex flows to match
At Dipli, every sales channel has its own model. Business customers pay by card, direct debit or bank transfer. Distribution partners (online stores, retailers, telecom operators) pay out lump sums, on their own schedule and by their own rules. Finally, trade-in programmes have no invoice to match against: the entries have to be built from Dipli's own trade-in data.
Matching a payment to its invoice is therefore anything but simple. A single payment can settle dozens of invoices and include credit notes, commissions or fees withheld by the partner. And the breakdown sits in scattered documents: statements downloaded from each partner's portal, files received by email, payout summaries, each in a different format.
As the business grows, volumes rise every month. Enzo Makaïa, Head of Finance, had set himself a demanding goal: a receivables ledger that is always up to date, collections based on an accurate view of outstanding balances, and eventually a monthly close. True to Dipli's culture of automating everything that can be automated, he made an early choice: hand this work to AI agents rather than dedicate internal resources to it.
The tech team could stay focused on an ambitious product roadmap. Dipli handed its payment matching to Sanso in March 2025, at a time when few small businesses were trusting AI with their finance operations.
The solution: AI agents that match, a team that validates
Sanso plugs into Pennylane and handles payment matching end to end, without involving Dipli's tech team. The process follows five steps:
- Collect the data and supporting documents. The AI agents pull invoices, credit notes and bank transactions from Pennylane, along with each partner's payout statements, whether downloaded online or received by email, in whatever format.
- Extract the allocation keys. From these statements, the agents identify what each payout covers: orders paid, credit notes deducted, commissions and fees withheld.
- Match and post the entries. Each transaction is matched with its invoices and credit notes, even when a single payout bundles hundreds of them. Sanso generates the required entries: partner commissions and fees, trade-in transactions.
- Get human sign-off. The Sanso team reviews discrepancies and edge cases before anything is synced. Anything that needs a decision goes to Dipli's team, with an explanation of each discrepancy.
- Update Pennylane. Once validated, the matches and entries are synced to Pennylane, where each transaction shows as matched with its invoices.
Enzo Makaïa designed the setup with Sanso one sales channel at a time: online payments first, then distribution partners, then trade-in programmes. For each one, he defined Dipli's own matching rules, tested them on a first month of data, then applied them to every transaction. Each new partner or trade-in programme is now added the same way.
The results: up-to-date books that keep pace with growth
From the first month, 96% of payments were matched automatically, and the rate has held ever since.
A receivables ledger that is always current. A few months after launch, 95% of customer accounts were fully matched. The team relies on Pennylane to track outstanding balances and organise collections.
We finally have books that tell us what's going on with receivables.
A calmer year-end close. Every payment received is tied to its invoice: the accounts are clear, for the statutory auditor and for management alike.
At year-end, our statutory auditor has everything right in front of them.
Sales and invoicing processes under control. Beyond the books, systematic matching acts as a safeguard. An order delivered with no payment, or a payment received with no matching invoice, cannot go unnoticed: every discrepancy surfaces and Dipli's team deals with it right away. Management can be confident that its sales and invoicing processes work as intended.
A finance team focused on its expertise. Enzo and Andrea, who make up Dipli's entire finance team, spend their time on analysis, planning and management control, where their expertise matters most. Each new sales channel comes on board with no extra workload.
Hiring people to match payments would have meant growing the team with every new channel. With AI agents, we stay a team of two and keep our time for analysis and strategy.