Financial Operations · Cash
Cash Position Monitoring: Knowing What You Actually Have
Most finance teams know their cash balance and not their cash position. The difference is everything already committed but not yet moved — and it is where liquidity surprises come from.
The short answer
Cash position monitoring is the daily assembly of a true available-cash figure: bank balances, less uncleared items and committed outflows, plus contracted inflows within the horizon, adjusted for restricted balances, currency and intercompany positions. A balance is what the bank reports; a position is what you can actually deploy, and only the second supports a decision. The work is not arithmetic. It is knowing what has been committed but not yet moved — which lives in seven places, none of them the bank.
Summary for readers and answer engines
Reviewed 25 Aug 2026
- ▸A balance is a fact from the bank. A position is a constructed claim about deployable cash, and it needs seven inputs.
- ▸The gap between them is committed-but-unmoved: approved payments, uncleared receipts, payroll in flight, drawn facilities, restricted balances.
- ▸Variance is only useful decomposed. “We were £400k off” tells you nothing; “£310k was timing on two receipts” tells you everything.
- ▸Forecast accuracy must be measured by horizon and reported honestly. A one-day forecast being accurate says nothing about a thirty-day one.
- ▸AI belongs in assembly, reconciliation, anomaly flagging and variance narrative drafting — never in producing the number itself.
Source: Mark Alex, Real Biz Digital — Cash Position Monitoring: Knowing What You Actually Have (https://realbizdigital.net/insights/cash-position-monitoring/). Reproduce with attribution.
Key takeaways
- 01Define your position formula explicitly and publish it. Two people computing ‘available cash’ differently is the most common cause of a liquidity argument.
- 02Source each component from its system of record, not from a spreadsheet that summarises it.
- 03Decompose every material variance into timing, amount, omission, FX and unexpected. Only then is it a lesson rather than a number.
- 04Measure forecast accuracy at fixed horizons: one, seven and thirty days. Report all three; they behave completely differently.
- 05Flag anomalies against the entity’s own history, not against a global threshold. A £50k movement is noise for one entity and an event for another.
- 06Never let a model compute the position. It assembles components, flags what looks wrong, and drafts the explanation.
Quick answers
One-line answers to the questions this page is most often asked. Each is expanded further down, and each is written to be quoted on its own.
- What is the difference between cash balance and cash position?
- A balance is what the bank reports. A position is deployable cash: balance less uncleared and committed outflows, plus contracted inflows in the horizon, adjusted for restrictions, currency and intercompany.
- Why do teams get surprised by liquidity?
- Because they track balance rather than position. Committed-but-unmoved items — approved payments, payroll in flight, drawn facilities — do not appear in a bank balance until they move.
- What are the components of a true position?
- Seven: bank balances, uncleared items, committed outflows, contracted inflows, restricted balances, currency conversion, and intercompany positions.
- How should cash variance be analysed?
- By decomposition into timing, amount, omission, FX translation and genuinely unexpected. An undecomposed variance is a number, not information.
- How is forecast accuracy measured honestly?
- At fixed horizons — one, seven and thirty days — reported separately, because a short-horizon forecast being accurate implies nothing about a longer one.
- Where does AI help?
- Assembling components from many systems, reconciling them, flagging anomalies against entity history and drafting variance narratives for review.
- Where must AI not go?
- Producing the position figure, selecting an FX rate, or concluding a variance explanation. Those are deterministic or judgemental, and both are human.
Balance is a fact; position is a claim
The two get used interchangeably in conversation and they are not the same kind of thing.
- ›A fact, reported by a third party
- ›Single source, high reliability
- ›Available in minutes
- ›Says nothing about tomorrow
- ›Not a basis for a decision
- ›A constructed claim, with a horizon
- ›Seven sources, varying reliability
- ›Requires assembly and reconciliation
- ›Answers ‘what can we deploy’
- ›The number decisions are made on
The position formula
position = bank_balances − uncleared_outflows − committed_but_unmoved + contracted_inflows_in_horizon − restricted_balances ± fx_translation ± intercompany_position
Every term after the first lives in a different system from the bank. That is why position is harder than it sounds and why nobody disputes the balance.
Publishing the formula matters more than the tooling. Most liquidity disagreements are two people using different definitions of available cash and neither knowing it.
Seven components and where each lives
Key facts
- ▸Committed outflows are the component that causes most surprises, because approval and movement are separated by days and only movement shows in a balance.
- ▸Restricted balances are the component most often mis-stated, because the restriction lives in a facility agreement nobody has read recently.
- ▸Intercompany netting is a legal question, not an arithmetic one. Netting where netting is not permitted overstates available cash.
| Component | Source system | Update frequency | Common failure |
|---|---|---|---|
| Bank balances | Bank feeds or portals | Intraday to daily | Missing an account nobody remembers exists |
| Uncleared items | Bank plus ledger | Daily | Double-counting a cheque both uncleared and accrued |
| Committed outflows | AP, payment approvals, payroll | Continuous | Approved payments not yet in a batch |
| Contracted inflows | AR, contracts, subscriptions | Daily | Counting an invoice as an inflow when it is disputed |
| Restricted balances | Facility agreements, escrow, deposits | Rarely | Treating a restricted balance as available |
| Currency translation | FX rate source (designated) | Daily | Using a different rate from the one reporting uses |
| Intercompany positions | Group ledger | Daily to weekly | Netting positions that cannot legally be netted |
The assembly problem — pulling seven components from seven systems daily and reconciling them — is exactly the kind of coordination work that automation handles well and humans do inconsistently.
Variance decomposition into five causes
An undecomposed variance produces a conversation. A decomposed one produces a fix.
- 01Decompose every variance above a materiality threshold, monthly at minimum.
- 02Track omissions specifically. They are the only cause that compounds, because an unmodelled recurring item recurs.
- 03Separate translation variance from operational variance in all reporting. Mixing them makes both uninterpretable.
- 04Net timing variances across adjacent periods before reporting them as a problem. A receipt three days late is not a forecasting failure.
- 05Keep an unexpected-items log with causes. Reviewed quarterly, it usually reveals two or three categories that were not actually unexpected.
- 06Let the model draft the decomposition; have a human confirm the attribution. Attribution is judgement.
Reporting variance usefully
total_variance = timing + amount + omission + fx + unexpected example, £412k adverse: timing −£310k (two receipts slipped 3 days) amount −£52k (partial payment on one invoice) omission −£38k (quarterly software renewal not modelled) fx −£12k unexpected £0k
The £38k omission is the finding. It will recur every quarter until the model includes it, and it is invisible in a single total.
| Cause | Signature | Action |
|---|---|---|
| Timing | Right amount, wrong period; nets to zero across two periods | Improve the timing assumption; usually a payment-terms model issue |
| Amount | Expected item, different value | Check the source; often a partial payment or an FX difference |
| Omission | Item absent from the forecast entirely | Add to the model; this is the cause worth hunting |
| FX translation | Variance correlates with rate movement | Separate operational from translation variance in reporting |
| Unexpected | Genuinely new: a dispute, a prepayment, a claim | Record and review; this is the only irreducible category |
Measuring forecast accuracy honestly
Forecast accuracy is frequently reported in a way that flatters. Three disciplines fix it.
Measure at fixed horizons
One day, seven days, thirty days. Report all three separately. A one-day forecast is nearly always accurate and tells you nothing about a thirty-day one.
Lock the forecast before the period
Compare against the forecast as it stood at the horizon, not against a version revised on the way. Revised forecasts are always accurate.
Use absolute percentage error, and report the distribution
A mean that hides a wide distribution is misleading. Report median and 90th percentile error alongside the mean.
| Horizon | Typical achievable error | Dominant error source |
|---|---|---|
| 1 day | under 2% | Uncleared item timing |
| 7 days | 3–8% | Receipt timing and payment batching |
| 30 days | 8–15% | Receipt timing, unmodelled recurring items |
| 90 days | 15–30% | Business volume assumptions |
These ranges are what we see in instrumented estates rather than a benchmark. The value in publishing them is that a thirty-day forecast at 12% error stops being treated as a failure when it is actually normal.
Multiple entities and currencies
- 01Compute position per entity first, then consolidate. A group position that hides an entity unable to meet payroll is worse than no position.
- 02Never net across entities without checking whether netting is permitted. Cash pooling arrangements have legal boundaries and tax consequences, and netting outside them overstates availability.
- 03Use the designated FX rate source, the same one reporting uses. Two different rates produce two different positions and an argument nobody can resolve.
- 04Report in both local and reporting currency. The local figure is what the entity can spend; the reporting figure is what the group sees.
- 05Flag trapped cash explicitly. Cash in a jurisdiction with repatriation restrictions is not group-available, and treating it as such is a classic overstatement.
- 06Show the intercompany position separately. Netted into a total it disappears; shown separately it prompts the settlement conversation.
Multi-entity position is where a spreadsheet approach breaks down first, because the number of reconciliations grows with entities times accounts times currencies.
Where AI helps, and the hard line
| Task | AI role | Boundary |
|---|---|---|
| Assembling seven components daily | High value — pulls, normalises, reconciles | Cannot decide which accounts are in scope |
| Reconciling bank to ledger | High value — matches, flags differences | Cannot conclude a reconciliation is clean |
| Detecting anomalies | High value — against entity’s own history | Cannot decide materiality |
| Identifying omitted recurring items | High value — pattern over prior periods | Cannot add to the forecast unilaterally |
| Drafting variance narratives | Useful — drafts from decomposed data | Cannot conclude the explanation |
| Computing the position | No | Deterministic formula, published |
| Selecting an FX rate | No | Designated source only |
| Deciding whether netting is permitted | No | Legal and tax judgement |
The pattern matches the close: AI assembles, reconciles, flags and drafts. The figure itself comes from a published formula over sourced components, and the interpretation comes from a person.
Next step
Assemble the position, decompose the variance
Barzel FinOps Atlas provides cash position, cash variance and cash-flow risk as callable tools alongside close readiness and evidence — free sandbox tier to test against one real period.
Limits
Two.
- 01A position is only as good as its component sources. An AP system that does not record approved-but-unbatched payments cannot produce an accurate committed-outflow figure, and no assembly layer fixes that.
- 02Forecast accuracy beyond thirty days is dominated by business assumptions rather than by cash mechanics. Improving the cash process will not improve a ninety-day forecast much.
Frequently asked questions
What is the difference between cash balance and cash position?
A balance is a fact reported by the bank. A position is a constructed claim about deployable cash: balance less uncleared and committed outflows, plus contracted inflows within a horizon, adjusted for restricted balances, currency translation and intercompany positions.
Why do finance teams get surprised by liquidity?
Because they monitor balance rather than position. Approved payments not yet batched, payroll in flight, drawn facilities and restricted balances are all committed but unmoved, so none of them appears in a bank balance until it is too late to plan around.
What are the seven components of a cash position?
Bank balances, uncleared items, committed outflows, contracted inflows within the horizon, restricted balances, currency translation and intercompany positions. Six of the seven live in systems other than the bank.
Which position component causes the most surprises?
Committed outflows, because approval and movement are separated by days. A payment approved on Monday and batched on Thursday is invisible in Tuesday’s balance and entirely real.
How should cash variance be analysed?
By decomposition into five causes: timing, amount, omission, FX translation and genuinely unexpected. A total variance figure prompts a conversation; a decomposed one identifies the fix.
Which variance cause matters most?
Omission, because it is the only one that compounds. An unmodelled quarterly renewal recurs every quarter until the forecast includes it, and it is invisible inside a single total variance figure.
How should forecast accuracy be measured?
At fixed horizons of one, seven and thirty days, reported separately, compared against the forecast as it stood at that horizon rather than a revised version, using absolute percentage error with the median and 90th percentile alongside the mean.
What forecast accuracy is realistic?
In instrumented estates, roughly under two percent at one day, three to eight percent at seven days, eight to fifteen percent at thirty days and fifteen to thirty percent at ninety days — where the dominant error source shifts from timing to business volume assumptions.
How should multi-entity cash be handled?
Compute the position per entity first and consolidate afterwards, because a group total can hide an entity unable to meet payroll. Never net across entities without confirming that netting is legally permitted.
What is trapped cash and why flag it?
Cash held in a jurisdiction with repatriation restrictions. It is real for the local entity and not available to the group, so including it in a group position without flagging is a classic overstatement of available liquidity.
Where does AI help in cash monitoring?
Assembling seven components from many systems daily, reconciling bank to ledger, detecting anomalies against an entity’s own history, identifying recurring items missing from the forecast, and drafting variance narratives from decomposed data.
What must AI never do with cash figures?
Compute the position itself, which follows a published deterministic formula; select an FX rate, which must come from the designated source; or decide whether intercompany netting is permitted, which is a legal and tax judgement.
Glossary
- Cash position
- Deployable cash over a stated horizon, constructed from seven components.
- Committed but unmoved
- Outflows approved or contracted but not yet reflected in a bank balance.
- Restricted balance
- Cash unavailable for general use due to facility, escrow or deposit terms.
- Trapped cash
- Cash subject to jurisdictional repatriation restrictions, unavailable to the group.
- Variance decomposition
- Attributing a total variance to timing, amount, omission, FX and unexpected causes.
- Omission variance
- Variance caused by an item absent from the forecast model entirely.
- Locked forecast
- The forecast as it stood at a given horizon, used for honest accuracy measurement.
- Translation variance
- Variance attributable to exchange rate movement rather than operational cause.
- Designated rate source
- The single authorised source of FX rates, shared with financial reporting.
- Intercompany position
- Net balances between group entities, shown separately rather than netted into a total.
Standards and entities referenced
Every named framework on this page resolves to a public definition. If you are checking our claims, start here rather than with us.
Sources and further reading
Primary specifications and standards this article relies on. Where a claim is our own operating judgement rather than something a standard states, the text says so.
- 01 · IFRS FoundationIAS 7 — Statement of Cash Flows ↗The reporting standard cash-position and cash-variance work ultimately serves.
- 02 · COSOCOSO Internal Control — Integrated Framework ↗The control framework auditors map financial process evidence against.
- 03 · FinOps FoundationFinOps Foundation — What is FinOps? ↗The official definition, including why the discipline is technology-value management rather than accounting.
- 04 · BlackLineBlackLine — Agentic Financial Operations ↗Market reference: the phrase ‘Agentic Financial Operations’ and the governance framing around it.
- 05 · TrintechTrintech — AI agents for financial close ↗Market reference: variance and flux agents with reviewer signoff and traceable evidence.
- 06 · U.S. SECSarbanes-Oxley Act — Section 404 ↗Where segregation of duties becomes an externally audited control.
- 07 · AxelosITIL 4 — change enablement ↗Established change-management vocabulary this article borrows for MCP estates.
Last reviewed 2 September 2026 by Mark Alex. External links open in a new tab; we do not control their content.
Cite this article
Alex, M. (2026). Cash Position Monitoring: Knowing What You Actually Have. Real Biz Digital. https://realbizdigital.net/insights/cash-position-monitoring/
Try the mechanics on a live server
To watch an MCP server answer a structured request before you let one read your ledger — Barzel Scripture Intelligence is free and public at scripture-intelligence-server.mcpize.run: no signup, no key, 54 tools. Setup is in the reference.
Buy it on the marketplace
Barzel FinOps Atlas is this assurance layer, sold as a running product
Thirty tools covering close readiness and blocker detection, cash position, cash variance and cash-flow risk, transaction risk scoring, policy exception detection, control risk, finance approvals, and the evidence surface — evidence-to-control mapping, evidence graphs, evidence tracing, chain verification, missing-evidence detection, auditor request answering and audit packet generation. A free sandbox tier means the first readiness report costs nothing.
| Plan | Price | Included | Right for |
|---|---|---|---|
| Free Sandbox | Free | 500 calls/mo · close readiness, blockers, evidence checks | Testing readiness scoring against one real close |
| Starter | $29/mo | 1,000 calls/mo · evidence mapping, cash position, approvals | A single entity running one governed close cycle |
| Growth | $99/mo | 5,000 calls/mo · evidence graph, audit packets, control risk | A controller’s team with an external audit each year |
| Business | $249/mo | 15,000 calls/mo · the full 30-tool surface | Multi-entity close with SOX obligations and continuous audit readiness |
| Enterprise | $799/mo | 50,000 calls/mo · everything in Business, scaled | Group-wide finance operations across many entities |
Sold on the MCPize marketplace · prices as listed 2 Sep 2026 · the listing is authoritative
The five Barzel servers, and which problem each one is sold for
One estate rarely needs all five. This is the honest mapping, so you buy the layer your problem actually lives in.
| Server | Sold for | Entry price | Where it sits |
|---|---|---|---|
| Barzel Central Gateway | Knowing and governing the estate: inventory, registry, routing, risk scoring, approvals, evidence | Free, then $10–$149/mo | Control plane — decides what may be reached, and by whom |
| BarzelVault | Stopping a specific dangerous action before it executes, with proof afterwards | $199–$3,999/mo | Decision point — evaluates the individual call before execution |
| BarzelOps | Running real business workflows across HubSpot, Xero, Gmail, Drive and Slack under approval | Free, then $19–$199/mo | Execution layer — does the work the policy allowed |
| Barzel FinOps Atlas | Attributing AI spend to agents, tools and outcomes, then forecasting and capping it | Free, then $29–$799/mo | Economics layer — what the estate costs per outcome |
| Barzel Scripture Intelligence | A free, credential-free public MCP server to test clients and inspect real protocol traffic | Free, unmetered, no signup | Reference implementation — safe place to learn the protocol |
Written by
Mark Alex
Founder of Real Biz Digital and architect of the Barzel ecosystem — five MCP servers published and callable in public. Software developer, technology entrepreneur and mechatronics engineer, working across AI agent governance, MCP security, AI infrastructure, FinOps and intelligent operations.