Stage 3 · Automate

Repetitive work to the agents. Signatures to people.

LABNEXUS is DaSP’s cloud platform for accredited laboratories. Its AI agents reduce repetitive work — with AICertus™ technology, every action traced — and the signature stays with the person responsible. They are already in use in accredited testing laboratories.

Already in use
In accredited testing laboratories, on real processes.
Every action traced
With AICertus™ technology: who did what, when, on what basis.
Your LIMS stays
The agents work on top of the LIMS or LIS you have, or on ATLAS.
Flusso di lavoro con verifica umana: documento, agente, verifica dell'esperto, firma01Documentonorma, SOP, rapporto di prova02Agentelegge, confronta, propone03Verifical'esperto controlla04Firmala decisione resta umanase qualcosa non torna, l'agente riprova con le indicazioni dell'esperto
How it works

AI is already in laboratories, without governance: a generic tool open on someone’s PC, a procedure rewritten without leaving a trace, an answer taken at face value without verification. Banning it makes it invisible.

Governing it makes it verifiable.

The LABNEXUS agents work on a process chosen with the quality manager, on top of the LIMS you have. They read the laboratory’s documents and records, propose, and stop where a person is needed: verification and signature stay with the person responsible.

Every step is recorded: it is material for the audit, not a risk to hide.

−40%In an accredited testing laboratory, LABNEXUS agents cut by 40% the time spent checking new revisions of UNI EN standards and updating standard operating procedures — work every quality manager knows well.

Where

In an accredited testing laboratory, on a process every quality manager knows.

What

Checking new revisions of UNI EN standards and updating standard operating procedures.

Who signs

As before: the person responsible. The agent prepares, the person decides.

Three acts: read, propose, verify and sign.

Scroll: the flow follows the story. Under each act, the processes already in production.

Act one

Read.

The agent reads what the laboratory already has: standards, procedures, records, test reports. From the laboratory’s versioned Knowledge Base, never from generic data.

  • Standards and SOPs in the version in force
  • The laboratory’s records and test reports
  • It knows what it may read and what it may not: it is written in its program

Act two

Propose.

The agent compares and proposes: what changes in a procedure, what is missing in a record, where a deviation lies. It proposes, it does not decide.

  • Flags, prepares files, gets things ready
  • Every proposal carries the source and the version of the document
  • Below the confidence threshold it stops and asks

Act three

Verify and sign.

The expert checks and signs, as today. If something does not add up, the agent tries again with their instructions. Every step stays recorded: who did what, when, on what basis.

  • Verification stays with a person
  • The signature with the person responsible
  • The register is material for the audit
Revisions of standardsDiagram of the flow, not a screenshot
  1. 01Documentstandard, SOP, test report
  2. 02Agentreads, compares, proposes
  3. 03Verificationthe expert checks
  4. 04Signaturethe decision stays human

if something does not add up, the agent tries again with the expert’s instructions

Checks new revisions of the standards and flags what changes in the SOPs. The decision stays human.

One process at a time. Yours.

Choose the process that steals most of your time: the agent is built around it. Eight processes already in production in accredited laboratories.

Try it: the program changes, the responsibilities do not.

The process

The agent’s program

Checking revisions of the standards

Reads
The UNI EN standards in force and the new revisions, the laboratory’s SOPs.
Prepares
The report of what changes and where the SOPs need updating.
Stops
When a change cannot be attributed with certainty to a procedure.
Verifies
Whoever verifies in the laboratory today
Signs
Whoever signs in the laboratory today

The decision stays human.

How we work: three phases.

  1. 01

    Process analysis

    We map where time is lost and where the risk lies, together with the quality manager. The laboratory internal audit can come before it.

  2. 02

    Agents configured

    On the chosen process, on top of the laboratory’s existing LIMS or LIS, or on ATLAS. You do not change the LIMS.

  3. 03

    Integration and training

    People learn to work with the agents and to check them.

The SCALE™ method

Five steps, the same for every agent.

  1. Scan · read

    The laboratory’s documents and records.

  2. Connect · link

    Requirements and procedures, to one another.

  3. Assess · weigh

    The deviations, by severity and risk.

  4. Lead · guide

    The expert validates and decides.

  5. Evolve · monitor

    Over time, with alerts.

Three entry levels.

You can start small: the scope is defined on your process.

First level

Adopt

To begin.

  • Concrete, low-risk activities
  • One process, one agent, one person checking it

For those who want to see an agent at work

Second level

Build

For the laboratory that has already decided.

  • Vertical applications for the laboratory
  • Governed AI, under the expert’s control

For those bringing the agents into a whole process

Third level

Orchestrate

For the laboratory that wants to go all the way.

  • Agents that follow entire flows
  • Under human control, traced

For those governing several processes together

Every action traced.
Signatures to people.Human in the Loop, the laboratory’s Knowledge Base, traceability: with AICertus™ technology.

On top of your LIMS or LIS

The agents connect to the system you have, ATLAS or another LIMS or LIS, and work alongside it. You do not change the LIMS.

The Knowledge Base belongs to the laboratory

Standards, procedures and records in the version in force: the agent reads those, never generic data.

Your data does not feed public models

It stays within the laboratory’s perimeter, on the LABNEXUS platform: servers in Switzerland, one installation per laboratory.

Where the data is stored: the Foundations stage

Reference standards

Designed in alignment with ISO/IEC 42001, ISO/IEC 27001, ISO/IEC 27017 and ISO/IEC 20000-1, the EU AI Act and the NIST AI RMF. Data processing compliant with the GDPR.

AICertus™ technology

The questions we get asked

AI puts accreditation at risk

The agents do not sign and do not decide in the critical phases. Every action is traced and reviewable: it is material for the audit, not a risk.

We have to change software

No. The agents work on top of the LIMS or LIS you have, or on ATLAS.

Who is accountable if it makes a mistake?

Whoever verifies and signs, as today. The agent prepares, the person decides.

Our data ends up in a model

No. No training of public models: the data stays within the laboratory’s perimeter.

How long does it take?

You start from a single process, at the laboratory’s pace. The first result comes on the chosen process, before deciding the next move.

Bring a process that steals your time

In the demo you see an agent at work on our environment: it reads, proposes, stops where a signature is needed. In the pilot the agent takes on one of your processes, on your methods and your records, with the person checking it.

See the demoAsk for a pilot on your data

The demo: 30 minutes, on our environment, free. The pilot is a quoted activity: you start from a single process.

Back to the stage Automate