Skip to content

Systems audit · priced to your scope

One process,
off your team,
in 60 days.*

Kwestra designs and runs production workflow systems for fintech, insurance and high-volume back office. The machine is that system, on synthetic invoices. The first one stops at a lever only a person can pull. Approve it and the rest post on their own.

*Typical for one well-scoped process. Some take longer.

Audits from $2,500, priced to your scope and credited against the build. You keep the process map whether or not you hire us.

Sample · INV-10482 Held · needs a person
KWESTRA · CL-01000
  1. 01 Intake Tray and nip roller
  2. 02 Extract Reading bar
  3. 03 Match Gauges and 0.85 bar
  4. 04 Human gate Stop bar and lever
  5. 05 Post Platen and stacker

  • Vendor Northwind Supply Co. 99
  • Invoice no. INV-10482 98
  • Invoice date 12 Mar 2026 97
  • Amount due $4,280.00 94
  • PO reference Not on the page Held

Human gate · PO reference

No purchase-order number on the page, and 0.61 is under the 0.85 line. A person decides before anything posts.

  • 60 days First system in production
  • 100% Of consequential actions pass a human gate
  • Always Written quote before any build
  • Even if you walk Process map is yours to keep

What posts

The record, and the row that did not

Four fields from INV-10482 clear on their own. The purchase order does not, because it is not on the page. Until someone approves, amends or rejects it, the system of record is untouched.

Invoice · INV-10482 Held
Field Value Confidence Disposition
Vendor Northwind Supply Co. 99 Cleared
Invoice no. INV-10482 98 Cleared
Invoice date 12 Mar 2026 97 Cleared
Amount due $4,280.00 94 Cleared
PO reference Not on the page 61 Held for a person

Synthetic invoice constructed to illustrate the approval gate. Not a client record, not a real payable.

The real question

Three ways an automation
project dies.

You have probably seen at least one of these. They are the reason the budget conversation is hard, and they are what the engagement is designed around.

  1. Scope found out later

    The vendor pitch ran clean on ten documents. Production had four hundred edge cases nobody had written down, and the fixed price became a change order every fortnight.

    What we do instead

    We map the process with the people who actually run it before anyone quotes. The exceptions go in the scope document, by name, with a decision for each.

  2. The black box

    It automated the judgement as well as the typing. Now nobody can explain a denial to a regulator, and the team quietly stops trusting the output.

    What we do instead

    Every consequential action stops at a human gate. The person approving sees what the system saw and why it proposed what it did, and that trail is the audit record.

  3. The pilot that never shipped

    Eighteen months, three vendors, one very good dashboard, and the team still does the work by hand because the thing never reached production.

    What we do instead

    One process, in production, in about sixty days. Not a pilot, not a proof of concept. A narrow thing that runs, with the second process scoped off what we learned.

Scope

What we actually build

Four shapes of work. If the process you have in mind is not one of these, the audit will tell you that in week one rather than in month four.

Document and intake pipelines

Invoices, claims, applications, KYC packets, remittance advice. Anything that arrives as a file or an email and ends up retyped into a system of record.

  • Extraction with confidence scoring
  • Exception queue with a named owner
  • Write-back to the system of record

Agent workflows with approval gates

Triage, routing, drafting and first-pass decisions, where a person signs off anything consequential before it leaves the building.

  • Proposed action plus its reasoning
  • Approve, amend or reject, logged
  • Full trail for audit and dispute

Integration where there is no API

The thirty-year-old core system, the portal with no export, the spreadsheet three teams depend on. Automation that holds up when the vendor changes a screen.

  • Resilient, monitored connectors
  • Alerting that names what broke
  • Runbook your team can act on

We run the line, not just build it

If you do not want to own the operation, we staff and run it on the system we built, and you hold us to a throughput number instead of a headcount.

  • Staffed queue coverage
  • Throughput and accuracy reporting
  • Hand back the keys whenever you want

Document and intake pipelines

Invoices, claims, applications, KYC packets, remittance advice. Anything that arrives as a file or an email and ends up retyped into a system of record.

  • Extraction with confidence scoring
  • Exception queue with a named owner
  • Write-back to the system of record

Agent workflows with approval gates

Triage, routing, drafting and first-pass decisions, where a person signs off anything consequential before it leaves the building.

  • Proposed action plus its reasoning
  • Approve, amend or reject, logged
  • Full trail for audit and dispute

Integration where there is no API

The thirty-year-old core system, the portal with no export, the spreadsheet three teams depend on. Automation that holds up when the vendor changes a screen.

  • Resilient, monitored connectors
  • Alerting that names what broke
  • Runbook your team can act on

We run the line, not just build it

If you do not want to own the operation, we staff and run it on the system we built, and you hold us to a throughput number instead of a headcount.

  • Staffed queue coverage
  • Throughput and accuracy reporting
  • Hand back the keys whenever you want

Price your audit

Weigh the work.

Five answers you already know. The price, the weeks and the credit come from them, and the work order prints them. You end it holding a map of how your operation actually runs, worth having even if you never hire us to build anything.

Work order · SA-01 Priced to your scope
Processes in scopeEach one is mapped on its own.
1
Roughly. It moves the price a little.
5 people
Systems it touchesERP, inbox, portal, spreadsheet, bank site.
2
Sites that run it differentlyBranches, entities or teams with their own way.
1
Regulated dataPick any that apply.

Book now

$2,500

Regular price $3,125

Launch offer: 20% off when you book by Dec 31, 2026. Your quote is held for 48 hours.

Work order
SA-01 Starter
Duration
2 weeks
Credited in full
$2,500
  1. SA-01 Starter $2,500
  2. SA-02 Team $3,000 to $6,000
  3. SA-03 Department $6,500 to $12,000
  4. SA-04 Division $13,000 to $20,000

If we finish the map and conclude the process is not worth automating, we say so, you keep the map, and we do not send you a build quote to be polite.

At the readout you are holding

  • Process map Current state, timed, every handoff and every exception.
  • Exception register The edge cases named individually, each with a decision attached.
  • Scoped build and quote Fixed scope in writing: what is in, what is out, timeline, price.
  • An honest verdict If the process should not be automated, that goes in writing too.

How the weeks run (shown for the two-week Starter)

  1. Day 1–2 Kickoff We sit with the people who run the process and watch it happen once, end to end.
  2. Day 3–6 Map Every step, handoff and exception written down and timed. Including the ones that only live in someone’s head.
  3. Day 7–9 Instrument Volume, cycle time, rework and error rates pulled from your data, not from a vendor benchmark.
  4. Day 10–12 Design Where the human gate goes, what gets automated, and what is better left alone.
  5. Day 13–14 Readout We walk your team through the map, the register and the quote, and leave all three with you.
Your time (Starter)
About four hours, total
Obligation after
None
You keep
Everything

The work order lists what you entered. If the work turns out bigger, we tell you at kickoff and you choose: we stick to what you entered, or we quote the difference before doing it.

Skip the call. Kickoff within 2 business days. Full refund if we're not a fit.

Skip the call

Start the audit now.

Pay the audit price and we start. No discovery call, no proposal round, no waiting for a slot.

  1. 01 Pay your audit price By card or US bank account, through Stripe.
  2. 02 An engineer reads your brief You hear from them within one business day.
  3. 03 Kickoff within 2 business days Then the weeks on your work order to the process map and a written quote.

Full refund if we're not a fit. If we read your brief and decide we're not the right team for it, you get the fee back. Otherwise it is credited against any build that follows: Credited in full: $2,500

SA-01 2 weeks $2,500

Secure payment by Stripe, card or US bank account. Bank payments take 3 to 4 business days to clear. By paying you agree to our terms.

Order of magnitude

What a build costs

Bands, not quotes. The audit produces the actual number, in writing, against a scope that names its own exclusions. We would rather you disqualify us now than in week six.

One process, one system of record

A single intake or back-office line. Extraction, validation, a human gate, write-back.

$18k – $35k 4–6 weeks

Multi-step with integrations

Several systems, real exception handling, a legacy platform with no usable API in the chain.

$35k – $60k 6–10 weeks

We run the line for you

Build, then staffed operation against an agreed throughput and accuracy number.

Build + monthly Rolling, 30 days’ notice

What moves the number

  • How many exceptions the process genuinely has, as opposed to how many people think it has
  • Whether the systems in the chain have an API or have to be driven through a screen
  • How much of the decision has to pass a human gate, and what that reviewer needs to see
  • Whether anyone has written the rules down before, or we are the first to try

Objections

The questions
that matter

If yours is not here, it is a better use of the call than a form.

How is this different from buying an automation platform?

A platform sells you the capability and leaves the hard part with you. The hard part is not the tooling, it is knowing what your process really does, including the exceptions nobody documented. We do that part first and then build on whatever tooling fits, including tools you already pay for.

We tried automation before and it did not stick. Why would this?

Usually it did not stick for one of three reasons: the scope was discovered during the build instead of before it, the system automated judgement and lost the team’s trust, or it never actually reached production. The audit exists to kill the first one, the human gate handles the second, and committing to one narrow process in about sixty days handles the third.

Do we need clean data and documented processes before we start?

No. If your processes were already documented accurately you would not need the audit. Undocumented, inconsistent and partly tribal is the normal starting condition and it is what the two weeks are for.

What happens to the people currently doing this work?

That is your call and we will not pretend otherwise. What we will tell you honestly is which parts of the work the system can take and which parts still need a person, because the exception queue almost always needs one. Most clients redeploy rather than reduce, because the backlog was the problem, not the headcount.

Who owns the system when the engagement ends?

You do. Code, infrastructure, documentation and runbook. We will run it for you if you would rather not, but that is a service you choose, not a dependency we engineer in.

What if the audit concludes we should not automate it?

We tell you, in writing, and you keep the map and the register anyway. It happens, and it is a better outcome than a build that was never going to pay for itself. We would rather lose the build than be the vendor in your next cautionary story.

How quickly can you start?

We run a small number of audits at a time so that the engineer who maps your process is the one who builds it. Book a call and we will tell you the next honest start date rather than the one you want to hear.

Next step

Put one process
on rails.

How it works

Send it over, and an engineer calls you

Two minutes on the form is enough. The engineer who would run your audit reads it and sets up thirty minutes with you. You leave knowing whether the process is worth automating, whether or not you book anything.

  1. 01 You describe the process Plain words, on the form. No deck, no questionnaire.
  2. 02 An engineer replies Within one business day. We say so if it is not a fit.
  3. 03 Thirty minutes on a call Then an audit priced to your scope, from $2,500, credited against the build.

Send the detail first

A person replies within one business day. Not a sequence.

Start now

Skip the call. Book the audit at the price on your work order and kickoff happens within 2 business days. Full refund if we read your brief and decide we're not a fit.

Book at $2,500 now