Skip to content
VEYQON
Company · Mission/02

Make the record true, then let it act.

A business decision is only as good as the record it was taken from. Our mission is to give an organisation one record it can trust, and then let intelligence act on that record inside limits a person set — in that order, because the reverse produces confident answers drawn from a fractured source.

One mission·One vision·Four things we will not do

Mission

Give every operating business one record it can trust, and intelligence that can act on it without anyone losing control of what happens next.

Two halves, and the order matters. Without the first, the second is a machine for producing plausible answers from bad data. Without the second, you have a filing cabinet with a search box.

Vision

Software that runs the operation, rather than software the operation has to be run around.

The measure is not how much of the day is spent in the system. It is how much of the day the system removed — the reconciliation nobody has to do, the approval that did not need chasing, the question that already had an answer attached to it.

In practice/01

What follows from that, in practice.

Each of these is a design rule with something on the site that depends on it. They are not aspirations; if we broke one, a page would have to change.

One record, not a copy of oneModules share a ledger rather than syncing to each other. Agents read the record itself, not a vector snapshot of it taken last week.
Authority is granted, never assumedAn agent has a scope, a ceiling, and an escalation path. It cannot approve its own work, and nothing it does is invisible afterwards.
Every answer carries its sourceIf the system tells you a number, it can show you the documents the number came from. An unciteable answer is a defect, not a feature.
Open where it countsThe core is open source and the schema is documented, so the record survives us. We contribute general improvements upstream rather than holding them.
Ship something real, earlyOne genuine process in production in the first weeks. A programme that shows nothing until month eighteen is a programme nobody can correct.
The exit is part of the offerBulk export of anything you can see, documented before you sign. Retention should be earned every year rather than engineered once.
Refusals/02
Each one costs us something

Four things we will not do.

A mission statement that costs nothing is decoration. These each cost us something specific, which is how you can tell they are real.

Price per agentIt would put us on the wrong side of the table — we would be arguing for more agents rather than for fewer, better-scoped ones. Seats and capacity are priced instead.
Let an agent close its own loopNo agent approves its own work, releases its own payment, or signs off its own exception. There is no configuration flag that turns this off.
Sell a plan we will not publishIf a wave plan only works in a proposal and not on a public page, it is not a plan. Every programme page carries its wave table.
Quote a result we cannot attributeNo percentage on this site without a named customer behind it. Category pages describe mechanisms; case studies carry figures.
Failure conditions/03

How we would know we were wrong.

The useful version of a mission is one with a failure condition attached. These are ours.

The record is not trusted
Somebody still keeps a private spreadsheet to check the system against. If that spreadsheet survives go-live, we did not finish.
The agent is switched off
Not because it was wrong, but because nobody could tell what it would do next. That is a scoping failure, and it is ours.
The programme slipped quietly
A wave moved and the customer found out afterwards. Waves move; finding out afterwards is the defect.
The export was never tested
An exit right nobody has exercised is a claim, not a right. It should be run at least once during the programme.

Next step

The commitments, with the implementation attached.

Security, deployment, and pricing are where these stop being statements.