CASTEL STUDIO
Order a scan

Where AI is worth testing — and where it isn't.

You describe how the work actually gets done. You get a written report: which parts are worth automating, which aren't, and what to do first. Sometimes the answer is that AI isn't the tool — we'd rather write that than sell you something.

Order a scan

written report · 1-2 weeks · founding-client rates

AN HONEST READ

An honest read, not a sales pitch

1 · It can say "don't"

The scan may conclude that a workflow isn't worth automating, that a manual fix is cheaper, or that a tool you already pay for does the job. That verdict costs the same as any other.

2 · You keep the decision

The report is a decision aid. You remain the only decision-maker — nothing here is an automated decision about your business.

3 · The report stands alone

It's yours whether or not you buy anything else from us. No implementation is bundled, promised or implied.

That's the difference: we get paid for the read, not for the build.

TWO SIZES

Two sizes

Light — 490€

  • Up to 3 workflows.
  • Written report: worth it / not worth it, per workflow, with the reasoning.
  • No call — you send the notes, we send the report.
  • Turnaround 1-2 weeks.

Standard — 890€

  • Up to 8 workflows.
  • Everything in Light, plus:
  • A 60-minute call before we write, so we're reading your work and not our assumptions.
  • A prioritised roadmap — what to do first, second, and what to leave alone.
  • Turnaround 1-2 weeks.

More than 8 workflows, or several teams? That's a scoping call, not a scan — write to us.

WHO THIS IS FOR

Who this is for

Small teams that keep hearing they "should be using AI" and want a straight answer before spending on it.

You've been quoted for a build

and you want a second opinion from someone who isn't the one building it.

Too many candidates

Several processes look automatable. You need an order, not a list.

Burned before

A tool was bought, nobody used it. You want to know why before repeating it.

You suspect the answer is no

and you want that written down by someone who looked.

WHAT YOU GET

What you get

  • A written report, per workflow: worth testing, not worth it, or fix it without AI first
  • A "don't build this" section — the candidates we rejected, and why
  • What it would take: data, effort, and what could go wrong
  • Standard: a prioritised roadmap and the call it's based on
  • The report is yours to take anywhere — including to another studio

How we work

  1. 01You send process notes (a document, an email, bullet points — whatever you have)
  2. 02Standard: a 60-min call
  3. 03We read and write
  4. 04Report delivered in 1-2 weeks
  5. 05You decide. That's the end of the obligation.

SAMPLE DELIVERABLE

What the report looks like

A Quick Scan is not a slide deck of generic AI ideas. It is a written working document: verdicts, risks, rejected ideas and a practical first roadmap.

REPORT STRUCTURE

Typical sections

  1. 01Executive verdict: test, skip, or fix without AI first
  2. 02Workflow map: inputs, decision points, outputs and owners
  3. 03Automation candidates with effort, data needs and dependency notes
  4. 04Rejected candidates and the reason they should not be built now
  5. 05Next-step roadmap with the smallest useful pilot

RISK REGISTER

Example rows

WorkflowRiskResponse
Invoice triageMediumHuman approval before export
Client reply draftLowCited source + review queue
Refund decisionHighDo not automate; write a checklist

ROADMAP

First four weeks

  1. Week 1Clean one source of truthClient + Castel Studio
  2. Week 2Build a controlled draft workflowCastel Studio
  3. Week 3Run ten real cases with manual reviewClient operator
  4. Week 4Decide: keep manual, pilot, or stopClient

Illustrative sample, not a fake client report. A real scan is written from your processes and may conclude that the best next step is not to build.

PRICE & ACCEPTANCE TEST

Order your scan

490€ Light / 890€ Standard

Founding-client rates — reduced while we build the public portfolio. Limited slots.

The acceptance test

We agree in writing what we're scanning before you pay. If the report doesn't cover the workflows we agreed, we fix it or refund it. What we won't promise is a particular conclusion — that's the whole point.

A sentence or two is plenty — no confidential details at this stage, please.

We'll confirm scope and send an invoice before any work starts. Nothing is charged from this form.

FAQ

Common questions, honest answers

What if you conclude we shouldn't use AI?

Then that's the report, and it's the same price. It's the cheapest possible version of that discovery — the expensive version is finding out after the build.

Do we have to buy the implementation from you?

No. The report is a standalone deliverable and it's written to be readable by anyone — including a studio that isn't us. If you do go ahead with us, the scan becomes the implementation brief.

What do you need from us?

However you already describe your work: a document, a process note, an email thread, bullet points. If it doesn't exist in writing, the Standard call is the faster path.

Is this an audit in the legal sense?

No. It's a technical read of your workflows by an engineer. It carries no certification and no legal standing.

Light or Standard?

Under three workflows and you can describe them in writing: Light. More than that, or the description would take longer to write than to say: Standard.

What this is not

  • Not a certification or a legal audit.
  • Not a promise of savings, ROI or a time figure — we haven't measured your business, and neither has anyone who quotes you one.
  • Not an automated decision — a person reads and writes it, and you decide.
  • Not a sales call in disguise — it's a paid deliverable, and "don't build it" is a valid outcome.

Order a scan

Order a scan

The report is a decision aid, not a certification and not a promise of a particular outcome. You stay the decision-maker.