Company
What Spark Is Building
Software can now read an invoice, check a stock level and decide what should happen next. In a growing number of settings, it can be trusted to carry out the task. Yet in the markets we care about, much of everyday commerce still lives in banknotes, delivery rounds and reconciliations done by hand at the end of the day.
Spark, formerly Broco, starts from that gap. We are building a payment network for cash-driven economies, beginning with logistics. The starting point is small and operational. The destination is not: an economy whose flows are visible, usable and programmable, and open to the software that will increasingly act on them. The path has four steps:
Step 1: digitize the cash.
Step 2: let payments circulate.
Step 3: add intelligence.
Step 4: open the network to software that can act.
Why cash-driven economies stay that way
Cash is usually described as a habit. It is closer to a chain of dependencies:
A customer pays a courier at the door.
The courier hands the money to a delivery company.
The delivery company settles with the merchant.
The merchant needs cash to pay a wholesaler, and the wholesaler needs cash to pay the brand or supplier upstream.
Each actor can only go digital if the next one can be paid digitally.
That is why adding a wallet at one end rarely changes anything. A merchant with a digital balance still meets the first supplier who only accepts notes, and the balance goes back to cash. The chain changes only when money can move from one participant to the next without leaving the system.
The chain is also hard to see. Today, three things sit in different places:
The order information
The responsibility for the cash
The availability of the funds
A collection can be recorded at the door before the notes are counted or handed over, so knowing that a payment happened is not the same as having funds ready to use. Merchants wait to be settled, delivery companies spend their evenings reconciling, and suppliers wait to be paid in turn.
Logistics is the way in, not the limit
Delivery already sits at the junction of these actors. It meets the customer, carries the merchant’s order and, where cash on delivery is common, collects the payment. Few other activities touch so many links of the chain in a single operation. That makes the collection at the door the natural first thing to digitize.
It also forces precision about what has to be built. Collecting a payment does not make it digital money. Several steps are involved, and they are distinct:
Step 1: record the collection against the order.
Step 2: reconcile it with what was due.
Step 3: take the cash into the network.
Step 4: settle the merchant digitally.
Each step has to hold before the next can be trusted. We are building that sequence, and how each step is carried out will depend on the market and the partners involved.
Once the sequence works, merchant settlement stops being an endpoint and payments start to circulate. A merchant holding a digital balance can pay a supplier, or another participant, without converting it back to cash first. Every additional use makes a balance worth holding, and every participant makes the network more useful to the next. The same network could connect businesses of very different shapes, from established distributors to the informal commerce that already depends on these chains.

How the network connects: suppliers, wholesalers and shops linked by B2B collections and COD settlement, with Spark cash points taking cash in and out. Every new link gives a balance one more thing to do.
From a connected network to the agentic economy
When collections, settlements and payees are structured and connected, a network can do more than move money. It can start by understanding:
Follow collections as they happen.
Flag a gap between a delivery and its payment.
Anticipate when a merchant will be short of cash before a supplier payment falls due.
Here, software observes and recommends, and people decide.
The next step is to act. This is where agentic AI comes in: software that does not only answer questions but takes steps toward a goal on someone’s behalf. In an agentic economy, such software takes part in commerce by reconciling accounts, preparing settlements and paying invoices. For that, an agent needs more than intelligence. It needs authorized access to structured flows, and a means of payment it can execute.
Here are two examples of what we aim to make possible:
A merchant sets a rule: pay this supplier once the delivery is confirmed and the funds are available. An agent prepares the payment, checks that the conditions hold, and executes it within limits the merchant has set.
A delivery company asks an agent to match the day’s collections to open orders, and to put only the exceptions in front of a person.
None of this works on cash alone. An agent cannot act on a banknote in a courier’s bag. Software can track a cash transaction. Acting on the funds requires a further connection: a digital balance and an authorized way to move it. Connecting physical transactions to payments that software can execute is strategic infrastructure, because it shapes which economies agents can act in. Spark intends to build that connection in its markets, so that cash-driven economies can take part instead of being left outside.
Why we are called Spark
A spark appears when something static starts to move. For us, it is the moment a payment collected in cash becomes digital and usable: first a settlement, then the next payment, then something that can be understood, and eventually something that software can act on, with permission.
Spark starts with the cash collected at delivery, and we want to build the infrastructure that lets cash-driven economies take part in the age of AI agents. The name carries both ends of that ambition: a single payment set in motion at the door, and a whole economy that can act once it is connected.



