Straight-through rate
The share of documents that reach their destination with nobody correcting them.
Automation projects go wrong in two ways: nobody knows what it will cost, and nobody agrees what finished looks like. We fix both before any code is written — one stream at a time, with acceptance measured in numbers.
the audit, free, no commitment
the pilot on one real document stream
agreed after the audit, not billed by the hour
Each one ends with something you can hold: a written scope, a working stream, a monthly report.
We map one document or enquiry stream end to end: where it arrives, who touches it, how many there are, which systems it has to reach and what usually goes wrong. You get a written scope, the three numbers the pilot will be judged by, and a fixed pilot price.
You get: scope + price on paperOne stream goes live on your real documents, not samples. Exceptions route to a person from day one. At the end we compare the agreed numbers with the baseline from the audit — if the stream misses them, we say so rather than declaring success.
You get: one stream running + measured resultThe remaining streams follow the proven pattern, one at a time, each with its own small scope. Rules stay editable by your team, and a monthly report shows volumes, straight-through rate and exceptions.
You get: the rest of the flow + supportAgreed at the audit, measured during the pilot. No percentages are promised before we have seen your documents.
The share of documents that reach their destination with nobody correcting them.
Time from arrival to posted, against the manual baseline we measure at the audit.
How often a person is needed, and for which reasons — the list drives the next round of rules.
Small, but real. A pilot stalls when these are missing, so we ask up front.
Someone who can say how a document should be coded and can approve the scope. Usually the accountant or operations lead, an hour or two a week during the pilot.
A test or limited account in the accounting or CRM system, and access to the inbox or channel the documents arrive in. We work with whatever access level your policy allows.
A month of genuine documents, including the ugly ones. Clean samples make a demo that fails in production.
How suppliers map to accounts, which cases must never be automatic. If it only exists in someone's head, the audit is where we write it down.
The audit is free because it is also how we decide whether to take the work.
Three working days of our time. If we think the payback is not there, we tell you and there is no invoice and no pitch.
Agreed in writing after the audit, based on what we saw. Not hourly, so an awkward document format is our problem, not your overrun.
Each additional stream is scoped and priced the same way, so you can stop after any of them.
Support, rule changes and the monthly report. Systems change their APIs; that maintenance is part of the arrangement, not a surprise.
SayFocus is a small studio, not an agency with account layers. The person who runs your audit is the person who builds the pipeline and the person who answers when something breaks.
Everything we build sits in your accounts: your accounting system, your CRM, your mailbox. If we part ways, the rules and the documentation go with you and access is revoked on the last day — there is no hostage data.
Pipelines are assembled on mature no-code and low-code automation platforms, with AI models doing the reading, and everything is deployed inside your own accounts. That is deliberate: the workflows, the credentials and the data stay yours, and nothing here is a proprietary box you can only rent from us.
Which tools go into a given build is decided at the audit, chosen for the task rather than the task bent to fit a tool. If we part ways, access is revoked and the rules and documentation stay with you.
Because the price depends on things we cannot see from outside: how many formats arrive, how clean the master data is, what the accounting system's API allows. Three days of looking is cheaper for both of us than a guess with a safety margin baked in.
We say so, show where the documents fail and what it would take to fix. You are free to stop there — you have paid a fixed pilot price and know exactly where the flow breaks, which is worth having either way.
No. The whole approach is built around the system you already use. If it has an API we post into it; if it does not, we prepare the file or the queue it can import.
Typically three to four weeks from the first call: three days of audit, about two weeks of pilot, plus the time it takes to get access. Our own accounting-firm project reached its first production run about two and a half weeks after the start.
They live in a place your team can edit — a table your accountants change themselves, not a black box. Every change is visible, and the review queue shows why a document was held rather than just rejecting it.
Three working days, one stream mapped end to end, a fixed pilot price on paper. Free, and we will tell you if it is not worth automating.
Get a free automation audit