Why this business has a particular cash shape
Three quite different businesses answer to the name. A services or BPO operation where people do the work. A hybrid where software does the first pass and people handle exceptions. And a software vendor selling processing capability to others. Their cost structures differ, but all three share the same timing problem in different proportions.
Costs are continuous and paid quickly. Payroll runs weekly or fortnightly and does not wait. Cloud and compute bills arrive monthly. Neither is negotiable in the way a supplier invoice sometimes is.
Revenue is billed after delivery and collected slowly. The client base is enterprise: banks, insurers, healthcare systems, law firms, government agencies. They procure slowly, pay slowly, and set terms of 30, 60 or 90 days because they are larger than you. Your leverage on payment terms is limited and it is worth being realistic about that rather than treating it as a negotiation failure.
Volume is lumpier than it looks. A large backlog conversion project ends and takes its revenue with it. A client insources. Seasonal filing peaks pass. Per-document pricing means revenue falls the moment volume does, while the team you hired to handle that volume is still on the payroll.
Put those together and the characteristic failure is a profitable business that cannot make payroll in a month when a large client pays late and a project finishes early in the same fortnight.
The five things that actually move cash
DSO, and the gap between terms and behaviour
A rule of thumb treats DSO around 45 days or below as healthy, but it only means something relative to the terms you agreed. If contracts say 90 days, a target of 60 is arithmetic that does not work. Track the gap between contracted terms and actual behaviour per client, because that gap is the part you can influence.
Receivables concentration
Most operations have a handful of clients making up the majority of revenue. That is a commercial risk everybody understands and a financing constraint most people discover late.
The payroll and compute cadence
How often money leaves matters as much as how much. A fortnightly payroll against 60-day collections needs roughly four to five payroll cycles of working capital funded before the first invoice for that work is paid.
Contract structure
Per page, per FTE, or against a monthly minimum changes the volatility of the receipt more than its size. The biggest lever available, and the one that gets least attention.
Delivery location and currency
Offshore centres mean local payroll, local statutory payments, and revenue in one currency against costs in another. An adverse move can erase a year of margin improvement with nothing operational going wrong.
Contract terms are the biggest lever
Most operators try to fix cash flow through collections. Collections is the smallest lever. The terms you signed set the ceiling on what collections can achieve.
| Term | Why it matters to cash | What to push for |
|---|---|---|
| Payment terms | Directly sets the size of the gap you have to fund | Net 30 where you have any leverage; raise it at renewal rather than mid-contract, when you have none |
| Mobilisation or setup fee | Cash in before delivery, on exactly the projects that require hiring first | Charge for onboarding, configuration and template build. It is real work and it is defensible |
| Minimum volume commitment | Converts variable revenue into a floor you can staff against | A monthly minimum with true-up above it, rather than pure per-document pricing |
| Milestone or progress billing | Cash during a long project rather than a single invoice at the end | Defined milestones on backlog conversions, tied to volume delivered rather than dates |
| Rate indexation | Protects margin across multi-year agreements | An annual adjustment clause tied to a published index, agreed at signature |
| Late payment interest | Gives collections something to point at | Include it. It is rarely invoked and it changes the tone of the third reminder |
| Termination for convenience | The cash cliff nobody models until it happens | A notice period long enough to redeploy or release staff without carrying them unpaid |
| Volume ramp assumptions | Overstaffing against forecast volume that arrives late or never | Tie your hiring to actual volume, not to the client's projection, however confident they sound |
The mobilisation fee is the single most underused item on that list. Enterprise clients routinely pay setup fees to software vendors without complaint and resist them from service providers out of habit rather than principle.
Financing, and what each option really costs
| Option | How it works | The catch |
|---|---|---|
| Bank revolver or receivables line | A borrowing base calculated on eligible receivables, drawn as needed | Cheapest option, hardest to qualify for. Concentration limits usually cap or exclude receivables from any client above a set percentage of the book |
| Invoice factoring | Invoices sold to a factor, which advances a portion of face value, typically in the region of 70 to 90 percent, paying the balance on collection less a fee | Faster and easier to obtain than bank debt, and generally more expensive. Recourse versus non-recourse determines who carries the loss if the client does not pay |
| Equipment finance | Scanners and capture hardware financed against the asset | Matches the asset life sensibly, and becomes less relevant every year as the work moves to cloud infrastructure |
| Term debt | Fixed amount and schedule, usually for an acquisition or a facility | Requires stable earnings. Lumpy revenue makes financial covenants genuinely risky rather than theoretically risky |
| Growth equity | Capital against ownership, most available to software-led firms | Costs nothing in cash today and the most of anything over time |
The concentration point deserves emphasis because it catches people out. The client that makes your revenue look impressive is the same client that will limit how much you can borrow against it. A lender assessing a receivables line will look at what happens if your largest payer stops paying, and will size the facility accordingly. Diversifying the client base therefore improves your financing capacity as well as your risk profile, which is a better argument for doing it than the one usually made.
On factoring specifically: it is a legitimate tool in a business whose main asset is receivables from creditworthy enterprise clients, and it is not a sign of distress. It is more expensive than bank borrowing, so the question is whether the work it unlocks earns more than the fee. If factoring an invoice lets you take a contract you would otherwise decline, the arithmetic usually works. If it is funding an operating loss, it is postponing a decision.
The forecast that actually helps
A monthly forecast is the wrong instrument for this business, because payroll does not run monthly. Thirteen weeks, updated weekly, is the standard for a reason.
- Build receipts from the receivables ledger, not from revenue. Take open invoices client by client and place them in the week you expect payment, not the week they are due.
- Use each client's actual behaviour rather than their contracted terms. If a payer has run at 74 days for two years, they will run at 74 days next quarter. Forecasting them at their contractual 60 builds an error into every week.
- Put payroll in first and by exact date. Including the months with three fortnightly runs, which is where unplanned overdrafts come from.
- Include statutory and tax payments per delivery country. These are large, dated and non-negotiable, and they are the most commonly omitted line in offshore operations.
- Model a volume-down case, not just a payment-delay case. Your largest client halving volume with 30 days notice is a more realistic scenario than a general slowdown, and it produces a different answer.
Alongside DSO, track receivables concentration as a standing number. Those two together tell you more about the risk in the business than the profit and loss does.
What changes when delivery shifts to AI
Most operators in this sector are somewhere in this transition, and it changes the cash profile in ways that are not all favourable.
Gross margin usually improves. Fewer hours per document is the whole point, and the effect on profitability is real.
But the cost base gets less flexible. Labour scales down with volume, imperfectly and painfully, but it scales. Engineering headcount and platform cost do not. A business that has replaced variable labour with fixed development cost feels a volume drop harder than it used to, which means the volume commitments discussed above matter more after the transition, not less.
Compute becomes a real cost of sales line. Per-document inference cost varies by document type and by how much of the file needs processing. Knowing unit economics by document type stops being an accounting nicety and becomes a pricing input.
Clients will ask for the savings back. They know what the technology does to your cost. Cost-plus per-page pricing invites that conversation at every renewal. Pricing against the client's outcome, with indexation agreed at signature, holds up better.
And you carry both models for a while. The transition period, running the legacy delivery alongside the new one, is a cash cost that rarely appears in the business case.
A note on why we published this
Plainly: this is outside what Uptiq sells. We build AI agents for lending and financial services workflows, not treasury or working capital tools, and nothing here is a product pitch. The reason it sits on our site is that we work alongside a lot of firms whose delivery is document processing, and the point in the section above, that moving from labour to software changes the cash profile as well as the margin, is the part most business cases leave out. Firms making that transition sometimes embed classification and extraction rather than build it, which changes the fixed and variable split described here. That is the only overlap, and it is worth being straight about it.
Frequently asked questions
Why do document processing companies run into cash flow problems?
A timing mismatch built into the model. Costs are consumed continuously and paid on a short cycle, since payroll runs weekly or fortnightly and compute bills monthly. Revenue is billed after delivery and collected from enterprise clients on 30, 60 or 90 day terms. Add lumpy project volume and receivables concentrated in a few large payers, and a profitable business can still be unable to fund a fortnight.
Is invoice factoring a good idea for this kind of business?
It can be, and it is not inherently a sign of distress. The main asset is receivables from creditworthy enterprise clients, which is exactly what factoring is designed for. A factor advances a portion of invoice value, commonly in the region of 70 to 90 percent, and pays the balance on collection less a fee. It is generally more expensive than bank borrowing, so the test is whether the capital unlocks work that earns more than the fee. Funding an operating loss with it postpones a decision rather than solving anything. Check whether the arrangement is recourse or non-recourse, since that determines who bears a client default.
How do we reduce DSO when clients are much larger than us?
Accept that terms are mostly set by the larger party and work on the gap between terms and behaviour instead, which is usually a couple of weeks and is genuinely influenceable. Invoice on the day the work completes rather than in a monthly batch, make sure the invoice matches the purchase order exactly so it does not enter dispute, know the specific person who approves it, and escalate on a schedule rather than when somebody remembers. Raise terms at renewal, when you have leverage, rather than mid-contract when you have none.
Which contract terms most improve cash flow?
A mobilisation or setup fee, because it brings cash in before delivery on exactly the projects that require hiring first. A monthly minimum volume commitment, because it converts variable revenue into a floor you can staff against. Milestone billing on long projects. And a rate indexation clause agreed at signature. Payment terms matter, but they are the term you have least influence over, which is why the others deserve more attention than they get.
How does client concentration affect our ability to borrow?
Directly. A receivables facility is sized on a borrowing base of eligible invoices, and lenders apply concentration limits that cap or exclude receivables from any single client above a set share of the total. So the client that makes revenue look strong is often the one limiting the facility. Diversifying the client base raises borrowing capacity as well as reducing risk, which tends to be a more persuasive internal argument than the risk case alone.
How does moving to AI-based processing change the cash picture?
Margin usually improves and flexibility usually falls. Variable labour is replaced by fixed engineering and platform cost, so a volume decline hits harder than it did before, which makes minimum volume commitments more valuable after the transition rather than less. Compute becomes a genuine cost of sales line that varies by document type, so unit economics per type become a pricing input. Expect clients to seek the savings at renewal, and expect to carry both delivery models during the transition.
This article is general information about business finance in a particular sector and is not financial, accounting, tax or legal advice. We are not financial advisers. Financing terms, advance rates, fees and eligibility vary widely by provider, geography and the individual business, and any figures described here are general market descriptions rather than quotes. Discuss your own circumstances with your accountant, and read any financing agreement with your own counsel before signing.
