Questions funders ask

Scriptre is a workforce of specialized departments that MCA funders employ to run their operation: intake, financial analysis, underwriting, verification, communications, closing, portfolio monitoring and renewals.

Nora, your chief of staff, coordinates eight agents: Ledger handles financial analysis, Vera underwriting, Atlas verification and investigations, Maya communications, Clara closing, Rowan portfolio and renewals, Iris workforce performance, and Knox security. Ledger and Vera work from the statements and your policy; neither is a second calculator or decides on its own. Work you’ve authorized continues on its own. Your people make the consequential decisions.

It isn’t a funder, and it never moves money. Funding happens outside Scriptre, and a person in your organization records it.

It isn’t a CRM with AI features added, and it isn’t an AI that decides on its own. Approving, declining, overriding a decline, setting terms and funding always belong to a named person.

Underwriters and operations staff work deal by deal: they review the evidence, make the decisions and clear the Action Queue. Owners and executives supervise the whole operation: what the workforce is doing, the pipeline, what’s waiting on a person, and the funded portfolio.

Each person has a role (admin, underwriter or viewer) that decides what they can see and do. Brokers and merchants aren’t users. They submit deals and receive messages, and they never see your scoring, findings or the workforce’s reasoning.

Code computes the facts from the statements, such as deposit totals, balances, negative days, returned items, remittance burden and which months are complete, and checks them against your own policy. The verdict comes from those rules, not from an AI model.

A deal’s first verdict is almost always Approve, with an offer sized from the numbers, or Decline (or Hard decline). It can also be Needs review, Insufficient evidence or Unavailable. What the merchant or broker asked for is shown for information only and never changes the verdict. A Decline names the exact policy line that failed, for example the monthly revenue against your minimum. COUNTER comes only later, as your answer to a broker who asks for different terms after an approval.

The model handles judgment that needs context, such as reading an unusual transaction description or noticing a pattern worth a closer look. Its judgment is shown to you as a flag, with the evidence behind it.

A flag can send a deal to Needs review. It never quietly changes a number or a verdict. A new judgment check runs alongside the existing rules on known deals, without affecting any verdict, until its results have been reviewed.

COUNTER is your answer to a broker who asked for different terms after an approval, such as more money or a different term. It is never Scriptre’s first verdict: an offer is sized from the numbers, not from what was asked. When a broker asks for something different, a person on your team reviews the request and either approves what was asked or counters.

No. Suggested offers are always labelled “Suggested · not approved yet”. Only a person in your organization supplies real terms, and only those terms can reach the merchant.

No. Bank Verification is a findings engine. It compares the merchant’s connected bank accounts with the statements they sent and reports what it found, such as an undisclosed account or a recent change in cash flow. It never approves, counters or declines a deal.

It also never calls a merchant fraudulent. A possible problem is shown as an indicator, with its evidence, for a person to review.

That depends on your operating mode, which an admin sets in Settings › Workforce. Supervised: nothing runs without a person. Approved Automation, the recommended mode that new organizations start in: internal work runs on its own, such as analysis, Atlas investigations and the actions you confirm with Nora, and every message that leaves Scriptre is a draft a person approves. Advanced Autonomy: Maya may also send the routine messages and decline notices you set to Send automatically, and offer emails that carry exactly the terms a person approved.

Shadow is optional, to ease your team in: the agents work every deal in the background and nothing is sent, decided or put in your Action Queue, except a question when Scriptre can’t find what your team decided. Settings › Workforce › Onboarding shows where the agents agreed with your team.

For each department and action, Settings › Workforce shows what happens right now and why.

In every mode: the offer terms themselves, contracts, recording funding and starting a renewal. Approving, declining, overriding a decline and setting terms are always a named person’s decision. Decline notices follow your setting for them and your operating mode, like Maya’s other messages.

No agent can change its own permissions, its budget or your policy.

Yes. An authorized underwriter can override Scriptre’s Decline with a recorded reason. The override is tied to that exact analysis, so if the evidence changes, the deal needs a fresh review.

We don’t publish an accuracy figure. Every number on a deal is computed from the statements and links back to the transactions behind it, so your team can check it. Where accuracy is measured, it comes from your own records and the corrections your team makes, never from estimates.

A technical failure is never a decline. If a statement can’t be read or a service isn’t responding, the deal shows Unavailable or Insufficient evidence and says what would fix it, such as re-uploading a page.

A failure stops only the work that depends on it. Everything else carries on.

Open the figure to see the transactions behind it. Your team can correct what was read; the correction is recorded with your name, and the original is kept.

You decide which reviewed corrections become evaluation cases, the examples the workforce is tested against. A single correction never becomes a rule on its own.

No. A figure Scriptre can’t establish is shown as unavailable or not established, never as zero or a guess. Policy thresholds are yours, and Scriptre never fills in a value you didn’t set.

Three plans, billed monthly: Starter, Growth and Scale, plus usage overages. Pricing is being finalized — request access. We’ll confirm what each plan includes, its price and the overage rates with you directly.

We haven’t measured typical onboarding time yet, so we don’t quote one.

The setup steps: invite your team and set their roles and two-factor sign-in; confirm who owns new deals; set up your policy in Settings › Underwriting rules; connect your intake email; and send your first deal.

Deals arrive by email at your Scriptre intake address, or forwarded to it, through broker upload links, or by upload: one file, many, or a zip. Coming soon: brokers submit through your broker portal.

Your servicing platform can send remittance events to Scriptre over a keyed connection, so Rowan can monitor funded deals. Connecting a specific platform is set up with us. Administrators can also connect a mailbox for communications.

Some connections are built but stay switched off until their outside service is in place: bank connections for Bank Verification, outside verification sources for Atlas, electronic signing, and voice transcription for Nora. Automations for your internal team’s tools aren’t available in the app yet. We don’t give dates for these.

No. Messages to brokers and merchants go only through Scriptre’s own submitter channel, under your sending settings. Automations for your internal team never contact a broker or merchant, and the submitter channel never carries internal notifications.

Eligible work keeps running in Scriptre’s services when the app is closed. Work that needs your desktop waits until it’s back. Every device is a window onto the same workforce and the same records.

Start with the Guide, or ask Nora about the work in front of you. For anything else, email support@scriptre.com. For privacy questions, email privacy@scriptre.com.