AI-built applications under control

One portal and user management

A web entry point for internal tools. Add and remove users and assign read or edit permissions per application so each person can access what they need.

Department workspaces

Separate areas within the same environment. Each department has its tools and access arrangements, with a shared visual identity and consistent behavior.

A common meaning for data

A shared data model and dictionary define customers, orders and projects. Changes affecting shared information require review and approval.

Bring order to growing collections of tools

This approach suits SMEs already building tools with AI or preparing to start: applications tied to a single computer, shared access without clear permissions, duplicated information or changes made directly to live tools. We define a foundation so each new application has clear responsibilities and agreed standards.

• Fewer isolated tools
• Clear ownership

Organize applications for order management, quotations, documentation or task and project tracking. Each department can develop tools around its own work. We assess which processes to simplify and what data they need, then prioritize applications by potential business value. These are illustrative examples; the specific applications are agreed with each business.

• Administration and operations
• Continuous improvement

For leadership

Visibility into the applications in use, their owners and the changes released to operations. A foundation for prioritizing improvements around real needs.

For departments

Autonomy to propose and evolve tools around daily work, using templates and guidelines rather than starting from scratch each time.

For users

A consistent experience across applications, clear access arrangements and documented knowledge to support learning and continuity.

Guided development

Templates, components and guidelines adapted to the business’s data model, roles and visual identity. Authorized people and AI tools use these standards to build consistent applications.

Agents with defined roles

We configure support agents for building applications, preparing tests, reviewing code and documenting it. Their instructions reflect the standards agreed with the business. We agree which changes the team can handle and which require specialist technical support.

We assess integration with ERP, CRM and other systems so applications can work with consistent data. First we check official access methods, licensing, data quality and available permissions. We agree which information is shared and whether access is read-only or also includes writing. Each integration is assessed for feasibility before committing to its scope.

• ERP and CRM integration
• Shared data

Define and build

The requester describes the need and acceptance criteria. An authorized creator develops the change from the template and stores it in a separate workspace in the repository.

Test and review

Automated checks and technical review precede functional validation in an environment separate from production, using test or anonymized data. Failed checks block release. The requesting department confirms the tool meets the need.

Approve and deploy

The accountable approver authorizes release to production. Administration manages access and the rollback procedure if an incident occurs.

Access and sensitive information

Two-factor authentication and permissions by person, department and application. We define who can read or change data and manage secrets outside application code.

Traceability and recovery

Code versions, logs of relevant actions, off-server backups of data and documents, and a procedure for recovering versions when incidents occur.

Monitoring and risk levels

Monitoring and failure alerts, including synchronization failures. Read-only tools, management applications and tools handling critical data receive reviews proportionate to their risk.

Control over code and access

We plan to hand over code developed specifically for the business, with the repository and accounts under its control. We agree usage rights for shared and third-party components, dependencies and the access needed to manage the applications.

A clear path to continuity

We design the foundation using standard technologies and document deployment, recovery and maintenance. The aim is to support handover to another team when needed. Moving to a different hosting provider or maintainer must be assessed against the architecture, licences and dependencies of each implementation.

We identify a priority process and review existing tools. We agree scope and acceptance criteria, prepare the environment and adapt or build a pilot application. We validate the complete workflow before expanding it. Training and documentation prepare the people responsible to administer the foundation and bring in further applications.

• Pilot and training
• Tailored support

A dedicated implementation

We prepare an installation for your business and adapt it to the agreed processes, data and roles. The environment can run on cloud infrastructure contracted for the client; it does not necessarily require a server at your premises.

A defined initial investment

Implementation covers environment setup, adaptations, integrations and the pilot included in the agreed scope. We define deliverables and acceptance criteria before work begins.

Separate infrastructure and AI costs

Hosting and infrastructure services are budgeted separately from implementation. We also distinguish AI subscriptions used to build applications from AI consumption during application use, where applicable. We agree providers, account ownership and who manages payment.

Agreed ongoing services

Maintenance, support and new features are contracted around the business’s needs. This separation makes the initial investment and recurring costs visible and helps assess how they change with usage.

Does building with AI mean the application always uses AI?

AI-assisted development uses these tools to write or review code. The resulting application may run without AI during use. If it includes text generation, document interpretation or other AI features, we agree data requirements, costs and oversight. The platform manages the application lifecycle in both cases.

Can we bring existing tools?

We review code, dependencies and access arrangements to assess adaptation. An existing application can become the pilot if it meets the agreed technical and functional criteria.

How do we add departments?

New areas reuse the portal, access system, templates and release workflow. Owners and permissions are assigned; new shared data and external connections are assessed before inclusion.

What is delivered at the end of implementation?

Depending on the agreed scope, deliverables include testing and production environments, the pilot application, repository and relevant code, templates, guidelines and documentation. Handover includes training for the people responsible. Deliverables and acceptance criteria are defined in the proposal.