# The data problem does not disappear when a better model arrives.

> When AI answers are unreliable, the problem may be missing, outdated or inaccessible information. Start with the sources your first useful task needs, rather than trying to fix all company data at once.

Canonical URL: https://marmation.com/solutions/data-readiness-for-ai/

## The decision

Is the information reliable enough to support this business decision?

Start with one important set of records. Establish which source wins a disagreement, who may use it, which version is current and how deleted or interrupted imports are handled before adding more data.

## The same policy exists in several conflicting versions

A company wants a knowledge assistant across five repositories. Documents have duplicate versions, inconsistent owners and policy updates that arrive by email. The team expects indexing to resolve the ambiguity.

## Search cannot decide which source the business trusts

Search can retrieve two contradictory documents perfectly. It cannot decide which is authoritative without business logic or source ownership. Bad records in downstream systems can make intelligent actions unsafe.

## Give each source an owner, access rules and a traceable history

List the sources, their owners, access rules and update patterns. Identify duplicates, missing values and disagreements before choosing what to index. Keep a reference dataset for testing. Every transformation should preserve where a value came from (its provenance), and every import should retain the record's history (its lineage), so a reviewer can trace an answer back to the source and recognize uncertainty.

## The source preparation and ongoing checks to agree

- Source inventory and data contracts with owners.
- Permission and classification mapping before ingestion.
- Quality, deduplication, versioning and lineage strategy.
- Retrieval and action data readiness checks plus monitoring.


## Some sources should stay outside the AI application

Not every source belongs in a new AI platform. Sometimes a better system of record or knowledge management policy is the right initial investment.

## Try changed permissions, deleted records and interrupted imports

Test changed permissions, stale revisions, deleted documents, duplicate identifiers, corrupt records and restart/resume of ingestion. Confirm data flows comply with approved policies.

## Begin with one information domain you can verify

Begin with one important information domain and measure readiness before scaling. A limited reliable corpus is more valuable than an impressive but ungoverned index.

## Related reading

- [Company knowledge should be useful to the people allowed to use it.](https://marmation.com/solutions/enterprise-knowledge-systems/)
- [Retrieval cannot be allowed to widen a user's permissions.](https://marmation.com/expertise/permission-aware-rag/)
- [Give an agent a job. Give the system boundaries.](https://marmation.com/solutions/ai-agents-and-workflows/)