by Agentronics

AGX Finance · Proposed read-only planning evaluation

Could payment timing
reduce borrowing costs?

Ask Treasury and Accounts Payable to compare your current payment calendar with a coordinated alternative over 13 weeks. Use the same reviewed forecast and permitted supplier terms to see the effect on borrowing costs, discounts and cash.

Customer qualification comes first. Current evidence comes from synthetic studies and test systems. Your records, financial terms and source mapping need qualification before evaluation. Scope, effort, fees and data handling are agreed first.

Read-only planning · Customer-owned decisions

One planning comparison.
A visible trade-off.

1 entityUSD, with one cash account
13 weeksone reviewed forecast and cash minimum
2 plansyour payment calendar and a permitted alternative

Start with a nonconfidential conversation. Describe the payment decision, available records and restrictions. We check whether the supported model fits before agreeing a supervised comparison.

Your team gathers existing records, approves the forecast and permissions, and reviews the proposed payment changes. We estimate preparation, delivery effort and costs after scoping. A wider pilot is a separate decision.

The evaluation changes no accounting or banking records, sends no payment instructions and executes no borrowing. Your team decides whether to use the recommendations through its existing process.

01 / What to bring

Use the records
your team already has.

For one dated planning snapshot, prepare the following package after scope and data handling are agreed.

Supplier records
Open invoices, balances, due dates and settlements, with accounting control totals and evidence of complete coverage. The starting scope supports up to 500 open supplier invoices.
Payment permissions
Your documented payment calendar, permitted alternative dates and evidenced discount terms. Identify critical suppliers and any payments that must stay fixed.
Cash and forecast
Opening cash, independent bank records, restrictions and the minimum cash balance. Treasury approves 13 weekly receipt and fixed-payment forecasts, including essential obligations and payments outside the selected invoices.
Borrowing terms
If financing is included, supply permitted draw amounts and dates, interest, repayment terms and existing debt service. Material terms must fit the supported model or block that scope.
Named reviewers
Treasury owns liquidity, funding and the forecast. Accounts Payable confirms supplier terms and commercial acceptability. Controllership checks the financial records. Name the people who approve and review the comparison.

Customer-created exports can support initial review, subject to mapping and reconciliation checks. An ERP connection is not a prerequisite. No confidential figures or bank credentials are needed for the first discussion.

Which financial terms fit the initial scope?

The starting scope is one USD entity and cash account, weekly balances and whole-invoice payments. Both plans use the same information, obligations, cash minimums and financing access. The reference follows your documented payment calendar, takes the discounts that policy would capture and selects its cheapest permitted funding. AGX additionally coordinates payment dates within approved supplier terms.

Optional financing must fit explicit borrowing choices, fixed weekly simple interest and full repayment within the horizon. Existing debt service remains a fixed obligation. Material fees, covenants, rate changes or other unsupported terms require separately qualified support, agreed exclusion of financing, or a not-ready decision; they cannot be ignored.

The evaluation uses your approved forecast. Customer-trained forecasting, other currencies, group consolidation, new financial-model categories and financial execution are outside this scope. If the decision depends on a projected covenant, use the separately scoped covenant-aware planning evaluation.

Covenant verification remains our lead assurance offer. Planning is an optional follow-on built on reviewed constraints, with its own forecasts, permissions and approval. For a covenant-aware scope, see the combined planning evaluation. ABL review is complementary and separate; its availability is not connected to planning.

02 / Before sharing records

Agree the comparison.
Then begin the work.

The first discussion uses a nonconfidential description of the decision and available records. Agree qualification work, preparation, effort and fees before commitment.

  1. Confirm the decision. Agree the snapshot date, reference payment calendar, permitted choices, approved forecast, cash minimum and borrowing terms. No demonstration assumption becomes a customer permission.
  2. Qualify the records and model. Import and reconcile a sample through the agreed mapping and actual calculation workflow. Unresolved data gaps or unsupported material terms mean not ready.
  3. Approve data handling. Agree the environment, providers, access, retention, deletion and evidence export before sharing company data. Use customer-created exports or separately qualified read-only access. Direct production bank connectivity and turnkey deployment in your cloud are outside this scope.
  4. Define success. Agree the comparison, independent checks, decision pack, customer review effort, value test, timetable and all-in cost before seeing results.

Missing mapping, integration or financial support is separately scoped and qualified before evaluation begins. The assessment does not include open-ended development or unapproved additional usage charges.

03 / Your decision pack

See what changes.
Judge whether it helps.

The financial comparison
Show the reference and selected plan side by side: borrowing interest, supplier discounts, peak borrowing, lowest forecast cash balance and debt remaining. Report whether each plan meets the reviewed constraints.
The payment changes
Identify which invoices move, their payment dates and amounts, and the effect on discounts. Accounts Payable reviews whether the permitted changes are commercially usable.
The supporting record
Trace material payments and the tightest cash week to the reviewed records, forecast and permissions. Retain the exact approved version, calculations, verification and open questions.
The evaluation scorecard
Compare modeled benefit with preparation, review, assistance and agreed costs. Your team decides whether the result is useful enough to repeat. A calculated opportunity is separate from savings achieved.
If useful: extend to the wider planning pilot

The separately agreed four-week proposal begins after readiness and scope approval. Its proposed three-business-day readiness assessment starts only once a complete sample snapshot and required access approvals are available.

The wider scope covers three historical decision dates and three weekly current snapshots, each with a 13-week horizon. Historical inputs must have been available at their original cutoff; later outcomes are reserved for checking. Four weeks does not establish thirteen weeks of future performance.

Six paired comparisons, two paired sensitivities and four fixed-choice controls make up 20 evaluations. The sensitivities test delayed collection and a larger cash reserve. Each control fixes every payment and funding choice to the reference and must reproduce cost, cash and debt exactly.

Week 1

Reviewed starting point

Confirm source mapping, reconciliation, the approved planning basis and effort in your existing process.

Week 2

Compare the plans

Review historical cases and the first current comparison, including payment changes, discounts, interest, cash and borrowing.

Week 3

Test changed conditions

Review a fresh comparison and the two approved sensitivities. Make the cost of additional cash protection visible.

Week 4

Repeat and decide

Complete the third current comparison, control tests and evidence exports. Assess value, total effort and remaining production work.

A plan that cannot meet the reviewed limits is reported as infeasible. Missing or unsupported information stays visible and blocks the affected calculation. Every conclusion remains conditional on the approved forecast and permissions.

04 / Success criteria

Correct first.
Useful to your team.

Agree the criteria before seeing results. These are proposed evaluation criteria, not demonstrated customer outcomes or service guarantees. Correct calculations and approval controls must pass regardless of financial benefit.

Reconciled inputs
Complete declared invoice and bank coverage, with zero unexplained difference in cents. Missing or conflicting evidence blocks the affected calculation.
Checked comparison
Each accepted plan passes financial checks and separate verification of the saved calculation. Independent review finds zero unexplained monetary discrepancy. Agree control cases before kickoff; the wider pilot retains all four fixed-choice controls.
Enforced approval
No calculation succeeds without its current approval. Edits or calculation-version changes require reapproval. Reject stale approval, failed reconciliation, unsupported material terms and unauthorized access. Approval authorizes analysis only.
Visible trade-offs
State forecast dependence, cash headroom and supplier consequences. Your reviewers can trace the five largest payments and tightest cash week. Unsupported risk coverage remains unavailable; applicable shortfalls require review before reuse or rerun.
Measured usefulness
Judge recommendations against the same constraints and a commercially usable reference. Count customer preparation, AGX assistance, review time and agreed costs. Report first-use effort separately from repeat use.
Proposed targets for the wider pilot

Agree these targets in the signed wider scope before kickoff; they are not guaranteed results.

Repeatability: your finance owner completes two scheduled refreshes without an AGX engineer editing data or code. Count all assistance. All four fixed-choice controls reproduce reference cost, cash and debt exactly.

Response: at least 19 of the 20 planned evaluations finish with verified results within 60 seconds of admission to calculation. Report import, human review and display times separately, retaining failures and timeouts.

Choose one primary value test before results: supported incremental financial benefit of at least 3× the agreed all-in cost for the same covered 13-week period, or at least 50% less preparation and review effort across the three current cycles with a positive customer-approved cost case. Count preparation, assistance, onboarding and recurring costs. Agree cost allocations and commercially usable recommendations; do not switch tests after seeing results.

Report each snapshot separately. Do not add overlapping forecasts, annualize a single event or count borrowed principal as savings. An infeasible reference supports no savings claim. A correct zero-benefit result can pass financial checks and fail the commercial test.

05 / Production decision

Review the evidence.
Choose the next step.

1. Accept the pilot
Your sponsor signs the scorecard. Unresolved financial, authority or data-handling failures stop progression. A failed value test means stop or agree a new scope.
2. Qualify production analysis
Separately contract and test the real import path, access and roles, deployment, retention, monitoring, recovery and release rollback. Repeat financial and approval checks on the exact production release. Require customer security acceptance and an operating owner.
3. Operate and monitor
Agree support and service commitments from measured workload evidence. Monitor freshness, reconciliation, errors and outcomes. Changes to models, mappings or the engine receive the relevant renewed review.

These stages cover read-only decision support. Automatic recalculation requires separate opt-in qualification. Payment or borrowing submission requires a separate execution capability and authorization scope.

Even the wider pilot’s three historical snapshots and three weekly observations cannot establish full-year savings, future forecast accuracy or production reliability. The final recommendation retains those limits. The synthetic CRONUS decision study shows the current demonstration and its evidence.

Assess fit before sharing data

Bring one decision.
Agree the evaluation.

Discuss the decision, available records and required controls first. The fixed delivery fee is quoted after readiness assessment and agreed before kickoff. No company-data upload is required to begin the discussion.

Discuss your financial priority →

Read the study and evidence