Quantum Beetle began with a simple observation. Enterprise software had become very good at recording what happened, and remarkably bad at understanding what should happen next.
Everything we build is an attempt to close that gap.
01A generation of recording
An ERP is a magnificent recorder. Every order, receipt, payment and movement is written down, with a date and a name. For decades that was the goal, and it was the right one. A company that cannot record cannot be run.
What the record does not do is think. It will accept a fourth code for the same part as readily as the first. It will pay the same invoice twice if the number differs by a character. It holds everything and understands nothing.
02What understanding would mean
- Knowing that two records describe the same thing.
- Knowing that this payment looks like one already made.
- Knowing that the part being ordered is in a store three aisles away.
- Knowing which clause in which contract answers the question being asked.
03Why it is possible now
For a long time, the only thing that could read a free-text description and know what it meant was a person. That has changed. Systems can now read, classify and compare by meaning, and do it over every record instead of a sample. Combined with ordinary rules and a person's judgement where it counts, that is enough to make the record understand itself.
04What I am not claiming
I am not claiming a machine that runs a company. The decisions stay with people. What changes is that every decision is made with the facts present: what already exists, what looks wrong, what the document says. The system proposes and explains. A person decides.
05An enterprise that understands itself
The ambition is simple to state. Not another copilot. Not another dashboard. An enterprise that can finally understand itself: its materials, its money, its purchases, its documents. I think that is worth building carefully, over a long time, and I think it is best built inside the enterprise, where it can be owned.
In short
- Enterprise software records thoroughly and understands very little.
- Understanding means knowing when records match, payments repeat and stock already exists.
- It is now possible to read and compare by meaning across every record.
- People still decide; the system makes sure the facts are present.
Questions
What observation did Quantum Beetle begin with?
That enterprise software had become very good at recording what happened and remarkably bad at understanding what should happen next.
Does he think AI will run companies?
No. In his account the system proposes and explains, and a person decides.
What is the stated ambition?
To make enterprises capable of understanding themselves.
Sounds like your problem?
Tell us about it. We'll say honestly whether the swarm can help, and what it would take.
Read next
- AI should sit inside the business, not on top of itConviction
- One brain, multiple limbs: why I built the company this wayArchitecture
- The layer behind the dashboardProduct thinking
- Why your AI should run on your servers, not mineSovereignty
- One company should own the resultHow we work
- A small business deserves the same engineering as an enterpriseWho we build for