Frequently asked questions

Answers for the finance lead comparing cash-forecasting, AI chasing, and dispute escalation

Same positioning we use in the side-by-side comparisons — written for CFOs and controllers evaluating AI-driven AP+AR on a single ledger. Anything missing? Mention it in the demo form and we will add it here.

Cash forecasting

  • How accurate are your AI cash forecasts?

    We publish variance bands by horizon — typically under 5% at the 30-day view and under 10% at 90 days once the model has 60 days of your posted invoices and receipts behind it. Forecasts are rolled up from the same ledger that drives AP and AR, so you do not maintain a parallel spreadsheet model that drifts out of sync the moment a receipt lands late.

  • Does the forecast update in real time or only at month-end?

    Continuously. Every posted invoice, expected receipt, scheduled payment, and FX revaluation feeds the forecast the moment it lands — not on a batch job at month-end. Because the forecast is rolled up from the same ledger as AP and AR, the cash position reflects what is actually owed and what is actually expected, not a stale snapshot someone rebuilt last Friday.

  • Does it handle multi-currency and multi-entity?

    Yes — multi-entity is native, with FX revaluation and a consolidated cash view that mirrors the $10M–$100M revenue band we are built for. You can drill from the consolidated forecast into any single entity or currency in one click, and the roll-up stays consistent because every entity writes to the same ledger model.

AI chasing

  • Does the AI really call customers, or is it just email reminders?

    Yes — an AI voice agent plus email plus SMS escalation runs 24/7, configurable per customer and per segment. The voice agent uses the same tone, payment terms, and dispute context your AR team would — but it picks up the first call at 7am and the last one at 11pm without a queue, and it pauses the moment a customer flags a dispute.

  • Can we control the tone and the frequency?

    Yes — per-segment presets (early / firm / final) ship out of the box and every line of copy is editable. You can throttle cadence per customer tier, skip a customer outright while a dispute is flagged, and A/B test a softer first touch on a strategic account without touching the rest of the book.

  • What happens when a customer replies?

    Outreach pauses the moment a reply lands. The reply — email, SMS, or a voice-agent transcript — is routed to the right AR analyst with the full thread and the originating invoice attached, so the analyst steps in with context instead of cold context. Outreach resumes only after the analyst signs off, and every resume is reason-coded in the audit trail.

Dispute escalation

  • How are disputes routed?

    By reason code, owner role, and customer tier — disputes over a configured dollar or aging threshold land on the controller, not the AR clerk. Routing rules are editable in plain English and tested against historical disputes before they go live, so a policy change does not silently strand a queue.

  • Do my team see the full payment and conversation context?

    Yes — every dispute record carries the invoice, the full payment history, the chase log, and the email or voice thread that triggered the escalation. Nothing has to be re-asked from the customer, and nothing has to be re-stitched from an inbox. Resolution notes feed the matching model so a repeat reason on the same customer resolves faster the second time.

  • Is the dispute trail auditable?

    Tamper-evident log — every status change, override, and write-off is reason-coded and timestamped, exportable straight into the audit binder. We complete a SOC 2 Type II in our first year of GA; the current readiness pack (control matrix, sub-processor list, and DPA) is available under NDA.

Still deciding?

See Paymind AI side-by-side against what you already run.