Genome
The core every server shares: reasoning, memory, safety checks and monitoring. Improved by us, delivered as signed updates.
Negenic™ by MeruTech
Intelligence, redefined.
Private AI. Built around your business.
Negenic is the technology your company uses to grow a private AI for its own work: support desk, reporting, projects. Your AI, named by you, learns your business from every verified answer and builds its own skills, on a private server that is yours alone.
- Your AI, your name
- Your servers or our private hosting
- Learns only what is verified
- Your data, never shared
The genetics of private AI
One genome. Many species.
Every Negenic server starts from the same proven core, then expresses only the genes your need calls for. What it learns for you stays with you.
Genes
One pack per need: the connectors, playbooks, rules and tests for a domain. ERP support, reporting, project control, or yours.
Expression
Your server: the core plus your packs, tuned to your systems and your history, on a server that is yours alone.
Evolution
Every job is a lesson in waiting. Only lessons that pass your tests survive. Natural selection, with your approval.
Inside every server
Not just a model. A complete AI server.
A model on its own guesses. Negenic surrounds it with everything a professional relies on: tools, memory, checks, and a record of every step. Your people simply talk to it, under the name your company chooses.
The engine
A capable language model running on your hardware, with an adapter trained for your need.
The loop
It plans, investigates in your systems, checks its work, and only then answers. A senior consultant, not a chatbot.
Knowledge
Your documents, your ticket history and your vendors' manuals, found in seconds and cited in every answer.
Memory
It remembers the facts it verified, with the date and the evidence, and lets stale ones go.
Learning
Verified work becomes lessons and skills. It retrains on your server, on your schedule.
Tests
Every change must pass your own test set before it goes live. No silent regressions.
Monitoring
Quality, speed, cost and every action on one console, with alerts when anything drifts.
Governance
It acts as the signed-in user, with their rights. Changes to production need a named approval.
Advisor
It can consult an outside expert model with general questions only. Off unless you switch it on.
Ask an advisor
A second opinion. Nothing shared.
When your server is unsure, it can ask a linked outside model for advice. A privacy gate first rewrites the question in general terms. Names, figures, systems and documents never leave.
- Off by default
- Every question logged and visible
- Advice is checked in your systems before use
- Advisor text never trains your model
- 1The question, on your server
- 2The privacy gate
- 3What leaves
- 4Advice, verified on your server
Why does the time import for Claire Martin on ACME-PROD reject 312 rows worth €48,200 since 3 October?
Why would a time-entry import reject a worker's rows after their assignment changed?
Check whether the entries still point to the worker's old assignment, then reprocess them against the new one.
Checked in your system: the assignment changed on 3 October. 312 rows reprocessed after approval.
Evolution, with a gate
Better every week. Nothing changes without proof.
Your lessons stay on your server. They are never pooled with other clients.
- WorkEvery job is recorded, step by step.
- OutcomeYour people mark answers right or wrong; checks and tests add their verdict.
- LessonMistakes and successes become candidate lessons.
- TestEach candidate is replayed on your test set. Most don't survive.
- ApproveA named person signs off what is left.
- PromoteThe lesson goes live, dated and reversible.
Start from a pack
Start from a pack. Grow the rest.
ERP support and delivery
Oracle Fusion tickets diagnosed with evidence, BI Publisher reports built and tested, mass data updates run as safe, reversible jobs.
Reporting on any data
Ask in plain words and get a report where every figure is proven against its source.
Project portfolio
Planning, cost and approvals for businesses that run on projects.
Your need
Legal, procurement, engineering, customer service: start from the core and grow your own pack.
Runs where your data lives
Your walls or ours. Your server. Your AI.
In your data centre
A single server on your premises. Nothing leaves the building.
In your cloud
Installed in your own cloud environment and region, under your security rules.
Hosted by us
A private server for you alone, run and maintained by us in the region you choose. Never shared with another client.
Rented AI or your own?
| Rented AI | Negenic™ | |
|---|---|---|
| Who builds it | The vendor, for everyone | Your company, with its AI doing the work |
| Your data | Sent to someone else's servers | Stays on your private server |
| Your business | Starts from zero every time | Remembers what it verified |
| Improvement | When the vendor ships a new model | Every week, with your approval |
| Checking its work | Yours to build | Built in, before every answer |
| Cost | Per use, rising as you grow | A server you control |
| Proof | Take it on trust | Your test set and a full audit trail |
Start growing your own AI.
Tell us about your work. We'll show you how your AI grows into it.
Talk to us