Skip to content
Technology and data

Zero Trust and System Integration — Security Around Existing Software

Understand Zero Trust, least-privilege access and the controls needed when integrating an ERP with other systems. Assess integration and replacement on their merits.

Zero Trust is a security approach based on explicit access decisions and limited privileges. A user or device is not trusted merely because it is inside a corporate network. Identity, permissions and relevant context inform what access is allowed.

This is different from deciding whether to replace an ERP or build an integration around it. Both a new system and an integration with an existing system need appropriate security controls.

The principles behind Zero Trust

The details depend on the organisation, but useful design questions include:

  • How is the identity of a person, service or device established?
  • Which permissions are necessary for the intended task?
  • What context affects an access decision?
  • How are access, policy changes and failures recorded?
  • How can access be reviewed or withdrawn when circumstances change?

Least privilege limits what an account can do. Segmentation and explicit service permissions can reduce the paths available after a compromise. Monitoring helps investigate behaviour, but no architecture eliminates all security risk.

Integrating an existing ERP

An organisation may want to connect an ERP with stock, sales or reporting systems. Using supported APIs or exports can sometimes avoid changes to the ERP’s internal code. That is an integration choice, not a definition of Zero Trust.

Before designing the connection, establish what interfaces are available, which operations are permitted and who owns the data. Read-only access and write access have different consequences. Even a separate integration can overload a system, expose information or apply an incorrect update.

An illustrative stock-reconciliation workflow

Suppose an ERP and a warehouse system record stock movements separately. A proposed integration could compare records and present discrepancies for review.

The design would need to define:

  1. how products and locations are matched;
  2. when each source is considered current;
  3. which system is authoritative for each field;
  4. what happens during partial failures or duplicate messages;
  5. whether corrections require human approval;
  6. how changes can be traced and reversed where appropriate.

A pilot can test these rules on representative records. It should not be presented as a guarantee that stock errors disappear at launch.

Security controls for the connection

Use credentials with the permissions needed for the specific task and manage them through an appropriate secrets process. Avoid sharing a broad administrator account across unrelated integrations.

Validate incoming data and handle unexpected responses. Define request limits, retries and failure behaviour so that a failing dependency does not trigger repeated harmful actions. Logs should support investigation without collecting unnecessary sensitive data.

Access reviews, dependency updates, monitoring and recovery procedures remain necessary after deployment. A diagram labelled Zero Trust does not establish that the controls work in practice.

When integration or replacement may fit

An integration may be suitable when the existing system still supports the business and exposes usable interfaces. Replacement may deserve consideration when support has ended, requirements cannot be met or maintaining workarounds costs more than a planned change.

Compare the options using the same requirements. Include data migration, training, dependencies, operating costs and risks. Neither approach has a universal delivery time or cost advantage.

Evidence before rollout

Test permissions, data mapping, failure handling and recovery. Include unusual inputs and degraded dependencies, not only the happy path. Establish who can approve changes and who responds when the integration needs attention.

Observability helps teams investigate system behaviour. Process mining may help examine workflows when suitable event records exist. Each is a separate capability with its own requirements.

Working with QA10

Q-Dev can develop a defined integration or application to an agreed scope. Within Q-Journey, an audit can inform the software developed in the next stage. The choice to retain, extend or replace a system follows the requirements and available evidence; it is not predetermined by a security slogan.

Describe the process, systems and access constraints when you contact us. We can use that information to discuss the next step.

GLOSSARY

Terms in this article

Clear explanations of the terms you may want to revisit.

Explore the glossary
Back to the knowledge base

From understanding to action Q-Journey

From business processes to practical change

Explore the path from company AI and an audit to software and ongoing development.