What a task looks like

Seven real examples, with the price our own estimator gives each one. Every number below came out of the estimator on our homepage, on these exact descriptions — you can run them yourself and compare. None of them is a quote for your situation: your task gets its own number in about fifteen seconds.

They are here because “we can do any task” tells you nothing, and a concrete task with a concrete price tells you everything. Every one below: 10 working days, 50% on signature, 50% on acceptance, one page of scope you approve before anything starts.

Entry offer

A report that fills itself every Monday

960 · 8 standard-hours

Every Monday someone exports last week's orders from our shop admin and pastes them into a Google Sheet for the sales meeting. We want that sheet to fill itself.

You get: The sheet populating itself on schedule, in the column order you already use, with a log of each run — so you can see it happened rather than find out on Monday that it didn't.

Why 8 hours: One source, one destination, a fixed schedule, no judgement calls. The smallest thing worth buying — and it lands exactly on the entry offer price.

Stop retyping won deals into your invoicing tool

1,920 · 16 standard-hours

When a deal is marked won in our CRM, someone manually creates the customer in our invoicing tool and copies the contact details. About 40 times a month.

You get: A working connection between the two systems, triggering on the status you name, with field mapping you approve and a log of every record it creates.

Why 16 hours: Two systems means two APIs, two authentication schemes, and a mapping that has to be right in both directions when a field is missing. Writing to a spreadsheet is one destination you control; writing into someone else's invoicing tool is not.

300 supplier invoices a month into one sheet

2,880 · 24 standard-hours

Every month we receive about 300 supplier invoices as PDF by email, in three different layouts. Our bookkeeper retypes them.

You get: A pipeline that reads the inbox, extracts the fields from all three layouts, writes them to one sheet, and flags anything it could not read with confidence instead of guessing.

Why 24 hours: Three layouts means three sets of parsing rules and a test set for each. The cost is not in reading one invoice — it is in being right about the two hundredth.

An EU AI Act assessment you can hand to an auditor

1,440 · 12 standard-hours

We use an AI tool in hiring. Our lawyer asked whether it counts as a high-risk system and what documentation we are missing.

You get: A written assessment as a matrix — obligation, what it requires, what you have, what is missing, what closing the gap involves — with regulatory citations so your lawyer can check the reasoning rather than take our word.

Why 12 hours: Precise legal-technical mapping, not opinion. This is a compliance assessment, not legal advice and not certification: your lawyer signs off, not us.

Four years of sales history, read properly

1,440 · 12 standard-hours

We have four years of sales history in a CSV export, about 18,000 order lines. Which customers are worth keeping, which are quietly leaving, which products actually make money after discounts?

You get: A written report with tables you can check against your own totals — customers ranked by what they are actually worth, the ones whose buying pattern says they are drifting away before they say it, and product margin after discounts.

Why 12 hours: The analysis is bounded and the answer is a document. No integration, no dashboard to maintain. If you want it to run monthly, that is a second task.

A market-entry brief before you commit

1,440 · 12 standard-hours

We are considering entering the German market. Who are the five main competitors, how do they price, what does our category require, and what does entry realistically cost?

You get: A brief with five named competitors and their actual public pricing, the regulatory requirements for your specific category, and a cost-of-entry estimate with its assumptions stated so you can argue with them.

Why 12 hours: Desk research and synthesis. No interviews, no primary research, no proprietary model. Every claim carries its source, and where the source is weak the brief says so instead of rounding it into confidence.

An assistant that answers only from your own documents

4,800 · 40 standard-hours

Our team keeps asking questions already answered in our handbook, contracts and past project files — about 400 documents.

You get: An assistant your team can ask in plain language, which cites the document it used and says it does not know rather than inventing an answer. Your documents stay yours; they are not used to train anything.

Why 40 hours: The largest thing we take as a single fixed-price task. Most of the cost is not building it — it is testing that the citations are right and that it refuses to guess.

What we do not take on

Honesty is cheaper than a failed project: we do not take multi-year system replacements, work requiring people on your site, anything needing a licence we do not hold, or safety- and life-critical systems. Above roughly 80 standard-hours we stop quoting a single fixed price and propose stages instead — a promise not to move the price is worth nothing if the number behind it was a guess. We also give no legal, tax or accounting advice: where a task touches those, we produce the analysis and your own professional signs off.

Your task is not on this list?

That is the normal case. Describe it on the homepage and see the number — no account, no card, and nobody calls you.

Get a price →