IT automation that asks before it acts, and learns from every answer.

[NOME PRODOTTO] reads the requests that reach your helpdesk, prepares the action on your infrastructure, and runs it only after a technician approves. Every correction becomes knowledge for next time.

The product is still being built. We are looking for the first companies to run it in production.

Mark RossEmail to helpdesk@08:14

Hi, I can't log in since this morning, it says account locked. Thanks

User
m.ross Mark Ross, Accounting
Issue
Account locked out
Since
Today at 08:12

Current state in Active Directory

LockedOut
True
BadLogonCount
5
LastBadPasswordAttempt
08:12
PasswordExpired
False

Based on

“Locked accounts” procedure, including a technician's correction from 12 March.

Prepared action

Unlock-ADAccount -Identity m.ross

Waiting for a technician

An example. The buttons work: the decision is yours.

[NOME PRODOTTO] in one minute

The video only loads when you start it.

The same requests, every day.

Locked accounts, new users, group changes, address assignments. First-line requests take up much of an IT team's time, and each one needs the attention of someone who knows what they are doing.

How to handle them properly, in that particular company, is often written down nowhere. It lives in the heads of the people who work there, and it leaves when they do.

How it works

Five steps. The fourth always belongs to a person.

  1. A request comes in

    By email to the helpdesk mailbox, or written in plain language. The way anyone would write it, with no forms to fill in.

    “I can't log in since this morning, it says account locked”

  2. The system understands it

    It works out who is writing, what they need and where to act. It uses what it knows about your company, not generic rules.

    User: m.ross, Accounting

    Issue: account locked

  3. It prepares the action

    It writes the exact command to run and places it next to the current state of the system and the company procedure it is based on.

    LockedOut True

    Unlock-ADAccount -Identity m.ross

  4. A technician decides

    Approve, correct or reject. Without approval, nothing happens.

    Technician's correction: before unlocking, check whether the password has expired.

  5. It runs and learns

    It runs the approved action and logs who decided what. The technician's correction goes into the knowledge base: next time, the action arrives already corrected.

    “Locked accounts” procedure updated.

What it can do

Today it works on the systems an IT team touches most often. The scope is growing.

Available

  • Users, groups and accounts in Active Directory
  • DNS
  • DHCP

Coming

  • Business systems such as SAP
  • Firewall consoles
  • In-house applications

These are systems you can't manage from a command line. [NOME PRODOTTO] will operate them the way a technician would, in front of the screen, under the same rule: approval first, action second.

Why it is different

It learns from corrections.

Other systems retrieve past cases that look like the new one, but they don't know whether those cases went well. [NOME PRODOTTO] does, because a technician told it. Every correction updates the procedure, and next time the action arrives already corrected.

The more you use it, the less work it gives you.

locked-accounts.mdplain text file

              
              
              # Locked accounts
            
              
              
               
            
              
              
              - Read LockedOut and BadLogonCount.
            
              
              + 
              - Also check PasswordExpired: if the
            
              
              + 
                password has expired, propose a change
            
              
              + 
                at next logon.
            
              
              
              - Unlock with Unlock-ADAccount.
            
Lines added by a technician's correction, 12 March.

Everything stays in your network.

It installs on your servers. The only thing that leaves is the text of the request, sent to the AI model under your account, not ours. No copies of your data on third-party servers. The alternatives are almost always cloud-only.

The decision stays with you.

It isn't an autonomy level you configure: it is how the system is built. No action touches your infrastructure unless a person has authorised it. And before deciding, the technician sees the exact command and the real state of the system, not a description.

The knowledge is yours.
The knowledge base is made of text documents in an open format, readable and editable with standard tools. It stays yours, and stays useful even without us.
No ticketing system required.
Many alternatives assume a ticketing platform that plenty of companies don't have, or only half use. [NOME PRODOTTO] starts from the helpdesk mailbox, which is always there.
Every action is logged, and any tampering with the log shows.
Who decided what, when, on what basis, with what outcome. Log entries are chained together, so a later change is detected.
It knows what it can't undo.
Reversible actions can be undone with one click. Those that can't are flagged before approval, not after.

Security

Before going to market, [NOME PRODOTTO] underwent an independent security review.

A system that prepares actions on your infrastructure should be examined by someone who didn't build it.

About us

Nervor was founded by four people. They are the ones building [NOME PRODOTTO], and the ones who will answer you.

  • [NOME]

    [RUOLO]

    [BREVE BIOGRAFIA]

  • [NOME]

    [RUOLO]

    [BREVE BIOGRAFIA]

  • [NOME]

    [RUOLO]

    [BREVE BIOGRAFIA]

  • [NOME]

    [RUOLO]

    [BREVE BIOGRAFIA]

Apply for the pilot programme

We are looking for the first companies to run [NOME PRODOTTO] in production. If you are a reseller or an investor, use the same form.

You are

Optional. For example: which systems you manage and how requests reach you today.Optional. For example: which clients you serve and what you offer them today.Optional.

We use this data only to reply to you. Read the privacy notice