In nearly every critical information system, there are accounts that do not correspond to any human user. They are called technical accounts, service accounts, or sometimes simply “the database account” or “the batch account.” They are created to allow one component to communicate with another, a scheduled job to run overnight, or a connector to access an external resource.
They are everywhere. And in many organizations, nobody knows exactly how many there are, who created them, what permissions they actually hold, or when their credentials were last changed.
A problem that grows in silence
Technical accounts generally don’t cause problems at the time of creation. A developer needs a service to access a database: they create an account with the necessary permissions. A batch job must call a partner API: a key is generated. A monitoring tool needs to read logs: it gets read access.
Each of these decisions is reasonable in isolation. The problem is accumulation.
After a few years, a mid-sized IS can have dozens or even hundreds of technical accounts. Some are documented in a versioned configuration file. Others are hardcoded in a deployment script. Still others exist only in the memory of an operator who has since left the team. And a significant proportion hold permissions far exceeding what is actually needed, because at the time of creation it was simpler to grant broad access than to finely calibrate permissions.
Excessive permissions are the norm, not the exception
The most common pattern is the technical account created with admin rights “to move fast,” with the intention of restricting later. That restriction almost never happens.
On a banking IS, an audit revealed that a service account used by a nightly consolidation batch had write permissions on every table in the production database, when it only needed to read three tables and write to one. This account had existed for four years, with the same password, known to three people, one of whom had left the company.
This is not an extreme case. It is an ordinary one.
The principle of least privilege is universally recognized. Its implementation for technical accounts is universally neglected, because nobody wants to risk breaking a process by restricting permissions whose actual usage is not fully understood.
Credential rotation is the simplest maturity test
If you want to quickly assess the governance level of technical accounts in your organization, ask one simple question: when was the password or key for each service account last changed?
In many organizations, the answer is “never since creation” for a significant proportion of accounts. Passwords are stored in configuration files, sometimes in plaintext, sometimes in a secrets vault but with no rotation policy. API keys are generated once and distributed via email or internal messaging. Certificates expire and are renewed in a rush, with no established process.
The difficulty is not technical. Tools like HashiCorp Vault, AWS Secrets Manager or Azure Key Vault support automatic rotation. The difficulty is organizational: to implement rotation, you first need to know exactly which components use which credentials, and test that rotation doesn’t break anything. On an IS where this mapping doesn’t exist, rotation becomes a project in itself, and it gets postponed indefinitely.
Technical accounts are underestimated compromise vectors
From an attacker’s perspective, a compromised technical account is often more valuable than a compromised user account. It frequently has elevated permissions. It is not subject to multi-factor authentication. Its activity is not monitored the same way as a human account, because nobody expects unusual behavior from it. And its credential never changes, which leaves an unlimited exploitation window once obtained.
The SolarWinds incident in 2020 illustrated this reality at scale. One of the access vectors relied on exploiting service accounts with elevated permissions in Active Directory environments. More recently, several compromises in the healthcare sector exploited technical integration accounts between systems, whose credentials were stored in plaintext in configuration files accessible from the internal network.
The problem is not that these accounts exist. It is that they escape the controls applied to human accounts: no MFA, no behavioral monitoring, no automatic revocation, no periodic review.
Compliance systematically stumbles on this subject
During an ISO 27001 or SOC 2 audit, or a sector-specific compliance review, technical accounts are almost always a friction point.
The auditor requests an inventory of accounts with access to sensitive data. The team provides the list of user accounts, with their profiles and permissions. The auditor then asks: “And the service accounts?” The silence that follows is telling.
Producing a comprehensive inventory of technical accounts, with their effective permissions, rotation policy and logging, is an exercise that few organizations can accomplish without significant effort. Not because the information doesn’t exist, but because it is scattered across configuration files, directories, secrets vaults, deployment scripts and operators’ memories.
For organizations subject to DORA or NIS2, technical account governance is no longer just a best practice. It is a verifiable requirement, with consequences for non-compliance.
What needs to be put in place
Regaining control of technical accounts does not necessarily require a large-scale project. But it does require method and consistency.
The first step is inventory. Census all non-human accounts that exist in directories, databases, file systems, secrets vaults and configuration files. For each one, document the component that uses it, the effective permissions, the creation date, the date of last credential change, and the responsible owner.
The second step is cleanup. Identify orphaned accounts (those whose associated component no longer exists), accounts with excessive permissions, and accounts whose credentials have never been changed. Revoke or restrict, starting with the most obvious cases.
The third step is ongoing governance. Integrate the creation and review of technical accounts into the deployment process. Enforce a rotation policy, even if it starts long (90 days is a good starting point). And above all, monitor the activity of these accounts just as you would monitor human accounts, detecting unusual usage patterns.
Conclusion
Technical accounts are the invisible infrastructure of every critical IS. Poorly governed, they become compromise vectors, compliance obstacles and grey areas that nobody wants to touch.
Inventorying technical accounts, restoring their governance, integrating rotation into the pipelines: that’s the work. Digital trust and managed services for trust-critical systems.