One Person Company Invoicing: The Missing Layer in the OPC Stack
Provincial governments are backing one-person companies with AI employees. Customer service, sales quotes, order tracking — all automated. But who follows up when the invoice goes unpaid?
Guangdong just became the first provincial government in China to roll out comprehensive support for one-person companies. Hubei followed days later. Shanghai, Wuhan, and Qingdao are offering computing subsidies, special loans, and incubator space. The OPC movement has gone from Reddit curiosity to government policy in under four months.
The model works like this: one founder, multiple AI agents, zero human employees. Steven Li, a cosmetics solopreneur in Jiangsu, runs four AI employees — customer service, sales quotes, order tracking, and performance reports — for roughly $40 per month across two ChatGPT Plus subscriptions. Rest of World reports that Suzhou alone plans to cultivate 1,000 one-person enterprises by 2028.
The stack is filling in fast. Customer-facing chat. Lead qualification. Content generation. CRM updates. Project management. Order fulfillment tracking. Each role gets assigned to an agent, and the founder oversees instead of executes.
But there is a gap in the middle of this stack that almost nobody is talking about.
The invoice is sent. Then what?
Most OPC founders automate the top and bottom of the customer lifecycle. An AI agent handles the initial inquiry. Another generates the quote. A third tracks the order. But the moment the invoice goes out, the workflow reverts to the founder refreshing their bank account and sending awkward follow-up emails.
This is not a technology problem. The tools to send invoices exist. Stripe, Xero, QuickBooks, Wave, FreshBooks — the market for invoice creation is saturated. What is missing is the layer between "invoice sent" and "payment received."
That layer is accounts receivable automation. And it is the single highest-ROI automation a one-person company can deploy.
Why AR automation matters more for OPCs than traditional businesses
A traditional business with 50 employees has an accounts receivable clerk. Maybe a collections department. Someone whose job is to check payment status, send reminders, escalate overdue invoices, and reconcile payments against open balances.
A one-person company has the founder.
When the founder is also the customer service agent, the sales team, the operations manager, and the accountant, the follow-up email for an overdue invoice is the easiest task to deprioritize. It is not urgent. It is uncomfortable. It can wait until tomorrow.
Tomorrow becomes next week. Next week becomes "I will write it off."
The numbers are consistent across every study of small business cash flow: between 30% and 50% of invoices are paid late. Not because clients refuse to pay. Because nobody follows up with the right cadence, at the right tone, at the right time.
What an AR automation layer actually does
An accounts receivable agent sits between your invoicing tool and your accounting system. It does not replace either one. It handles the workflow between them:
Day 0: Invoice is created in Stripe (or whatever you use). The AR agent registers it and starts tracking.
Day 1: Payment confirmation check. If paid, mark resolved and sync to accounting. If unpaid, wait.
Day 3: First follow-up. Friendly tone. "Just confirming you received our invoice." This catches the legitimate "it went to spam" and "I forgot" cases.
Day 7: Second follow-up. Slightly more direct. References the specific invoice number, amount, and due date.
Day 14: Escalation. The tone shifts. The message explicitly states the invoice is overdue and asks if there is an issue.
Day 30+: Human-in-the-loop trigger. The agent flags the account for the founder to make a judgment call — is this a dispute, a cash flow problem on the client's end, or a bad debt?
At each step, the agent adjusts based on the client's behavior. If the client opened the invoice but did not pay, the follow-up message is different than if the invoice was never opened. If the client has a history of paying on day 25, the agent does not escalate at day 14.
This is not a drip email sequence. It is a stateful, context-aware collection workflow.
The math for a $40/month OPC
Steven Li's four AI employees cost $40/month. A cosmetics business selling to overseas customers almost certainly deals with net-30 or net-60 payment terms. If his monthly revenue is $5,000 and 35% of invoices are paid late — a conservative estimate for international B2B — that is $1,750 in cash sitting in limbo every month.
A single automated follow-up sequence that moves the average payment date from day 45 to day 20 does not add revenue. It adds cash flow. For a one-person company with no credit line and no employees to absorb payroll delays, cash flow is survival.
The cost of an AR automation agent is negligible next to the cost of a single unpaid invoice. If one $500 invoice that would have gone to write-off gets collected because an automated reminder went out on day 7, the tool has paid for itself for the year.
What to look for in an OPC invoicing stack
The OPC invoicing stack needs three components, and most founders only have two:
1. Invoice creation and delivery. Stripe Invoicing, FreshBooks, Wave, Xero — take your pick. This is solved. The invoice gets created and sent.
2. Accounting and reconciliation. QuickBooks, Xero, Wave — the payment lands, it gets categorized, tax reports get generated. Also solved.
3. Collection workflow. The layer between 1 and 2. This is where the follow-up happens, the payment status gets tracked, the reminders get sent, the escalation logic runs. This is the layer most OPCs are missing entirely.
AgentReceivable was built specifically for this third layer. It connects to Stripe for payment status, syncs with Xero or QuickBooks for reconciliation, and manages the entire collection workflow autonomously — from first reminder to human-in-the-loop escalation. It runs on OpenClaw, which means it fits into the same agent infrastructure that OPCs are already building.
For $19/month on the Pro plan, it is less than half the cost of Li's entire AI employee roster. And it handles the one workflow that directly impacts whether revenue stays theoretical or becomes cash in the bank.
The OPC stack is not complete without AR
The Chinese government is betting on one-person companies as an economic model. Guangdong, Hubei, Shanghai, Shenzhen — provincial and city governments are providing subsidies, incubator space, and computing resources. Qianhai launched an OPC Mavericks accelerator offering free office space, computing power, and seed funding.
This is not a temporary trend. The infrastructure is being built to support thousands of AI-powered solopreneurs.
But the stack these founders are assembling has a hole. Customer service is automated. Sales quotes are automated. Order tracking is automated. Performance reports are automated. The one workflow that directly determines whether the business has cash to operate — collecting payment on outstanding invoices — is still being done manually, if it is being done at all.
The next wave of OPC tooling will not be another chatbot or another project management agent. It will be the boring, high-leverage back-office automation that determines whether a one-person company stays in business: getting paid on time, every time.