# Bring AI into the environment you are responsible for.

> Bring AI into the systems you are responsible for, with a clear view of company data, employee access, running costs and who will support it when something changes.

Canonical URL: https://marmation.com/who-we-help/technology-and-ai-leaders/

The IT team rarely has the luxury of evaluating a model in isolation. It is responsible for identity, data, contracts, systems, security and what happens when a vendor or API changes.

## Pressure to deploy meets responsibility to operate

You may have multiple AI pilots, inconsistent data sources and pressure from business units to deploy faster. Some employees want to use consumer tools; others need specialized knowledge or workflow applications.

## One more assistant can create one more liability

A one-off chatbot can create more shadow IT. A blanket platform strategy may fail on unusual workflows. A self-hosted model may impose costs and maintenance your team has not planned.

## Compare the options against your actual restrictions

Choose a task and the access rules it must follow. Compare existing products with custom software, check whether the necessary sources and systems can be connected, then test the proposed approach on examples your team agrees are representative.

## Engineering around access, actions and operation

- Choosing between privately operated models and managed services, with a map of where company information would go.
- Company search and knowledge applications that obey each employee’s access rights.
- Agents allowed to use only defined tools, with a clear point where a person must approve a change.
- Connections to existing systems, quality and cost monitoring, and the documentation and access agreed for handover.

## Test the platform you already have

We do not recommend a custom retrieval-augmented generation (RAG) system, which finds source material for a model to use in an answer, when Microsoft, AWS, Glean or another approved tool solves the problem. Custom engineering needs a specific justification.

## Evidence your team will need after launch

Ask for a permissions test, a list of external services and licensed components, a maintenance and retirement plan, and tests of requests the system must refuse. The system must be maintainable after the first demonstration.

## Choose one valuable task with clear limits

Start with the most valuable constrained workflow, not a company-wide claim that every employee needs a new AI assistant.

## Related reading

- [Make AI work with the systems your business already runs.](https://marmation.com/solutions/enterprise-ai-systems/)
- [Private AI is not one architecture.](https://marmation.com/solutions/private-enterprise-ai/)
- [Retrieval cannot be allowed to widen a user's permissions.](https://marmation.com/expertise/permission-aware-rag/)