Changes verified before they go live
A new field, screen, or workflow is first validated by the platform, then saved as the next version. Reverting a change is one step, not another project with a schedule.
Platform
The application as a model.
On our platform, the application is a description of data, screens, and rules that the system understands in full. This means changes can be reviewed before going live, versioned, and reverted. Including the ones made by AI.
- Every element of the application is described and documented.
- Data in a PostgreSQL database, on your infrastructure.
- Changes visible immediately, with no compilation or maintenance windows.
Director · prompt
Add a "safety review date" field to the employee record, with a reminder.
EN
Entity: employee9 fields · Postgres
Validation OK +1
New field: safety review datetype: date · rule: reminder
Added v8
Model version 8AI agent · audited
Revert: 1 step modelentities · screens · rules — read livelayersplatform · customisations — separateApplication model
What follows from the application being described.
In classic IT, every change is a small project: a developer, tests, a deployment. The problem is not the people but the material they work with: an application made of code, where it is easy to break something without noticing. Here, the application is a model the platform understands in full. That gives three practical differences.
Updates without breaking customisations
Your customisations and the platform live in separate layers. The platform can be updated without touching what was built specifically for you. Salesforce rests on the same mechanism.
AI that does not guess
The agent does not dig through code. It reads the model and rules the platform understands in full, so every change it makes can be reviewed before it reaches users.
A change to a system should be a decision, not a project.
Data ownership
The data lives with you.
Every business application is screens and data. Data is worth asking about before the purchase, not when parting ways with the vendor. Our answer is simple: the data lives in your own PostgreSQL database, on your servers or in a cloud chosen by your security team. No licences and no fees for capacity.
- Ownership without asterisks. We maintain the database and build applications on it. You remain the owner, and you can take the data out at any time.
- AI directly on the data. The agent works straight on your database: it analyses and reports with no limits, intermediaries, or extra fees.
- The database AI knows best. PostgreSQL is the database best represented in the training data of language models, so they make the fewest mistakes on it.
A typical low-code platform
Salesforce · Power Apps
- Data on the vendor's servers, in a closed format
- Access through the vendor's API, with limits
- Fees for capacity and per user, every month
- AI access to the data billed separately
Our approach
DEX · PostgreSQL
- Data in your own database, in an open format
- Full access: your servers or a cloud of your choosing
- $0 in database licences, no fees for capacity
- AI works straight on your data, with no limits
System development
Changes visible immediately.
In classic technologies, an application is first programmed, then compiled and deployed. Every change, however small, travels that whole road. That is where the schedules, maintenance windows, and sign-off meetings around every fix come from.
Our platform
Here, the application is not compiled. The platform reads the model live, so a change in the model is visible in the running system immediately, and reverting it takes one step. For the same reason, AI can make changes safely: it works on a description of the application that can be reviewed, not on code that has to be read line by line.
Security and control
Control down to a single row.
Because the platform understands the whole application, it can guard things that loose code cannot: consistent permissions, a full history of changes, and continuity of operation whatever happens to the vendor.
Access down to a single row
Who sees which record is decided by a rule, not by accident. Permissions work consistently across every screen and every API, down to the level of a single row of data.
A full history of every change
Who changed what and when, in the model and in the data. Every change is versioned and traceable. No guessing and no digging through correspondence.
Platform code in escrow
The platform code goes into escrow. Even if our company disappeared from the market, you are left with a working system, access to the code, and full control.
Zerouser training after migration
Daysfor a change that used to be a project
$0in database licence fees
100%of logic migrated and documented
Access to your own data should not have a price list.
Let's talk about your system.
20 minutes. We will show the platform live and answer your IT team's questions. No obligations.