Semantic models carry business meaning
Google Cloud has made expanded semantic-model support in Knowledge Catalog generally available. The service can ingest definitions and relationships from Looker LookML, BigQuery graph and measure features, dbt MetricFlow, Snowflake Semantic Views, Databricks Metric Views and SAP Core Schema Notation. The goal is to give analytics tools and agents a shared representation of business meaning.
Database columns rarely explain themselves. A field called revenue may represent bookings, recognised revenue or a regional calculation, while a customer relationship can depend on several identifiers and time rules. Semantic models encode those definitions so a system can choose the right measure and join path instead of guessing from names or repeatedly embedding instructions in a prompt.
A context graph can reduce repeated prompt engineering
Knowledge Catalog brings the imported models into a context graph with technical metadata and governance information. An agent can use that graph to locate an approved definition, understand dependencies and cite the data asset behind an answer. This can make retrieval more efficient and reduce the chance that two teams obtain incompatible answers from the same question.
Google reports that grounding with semantic context produced nearly four times the accuracy of basic retrieval in an internal benchmark, with up to 33 per cent lower latency and 75 per cent fewer tokens. These results are useful indicators rather than universal guarantees. Performance will depend on model coverage, metadata quality, question type and how conflicts are resolved.
Conflicts and lineage still need governance
Cross-platform ingestion also exposes governance challenges. Two systems may define the same metric differently, an imported model may be stale, or a technically valid relationship may not be approved for a particular audience. Owners need precedence rules, freshness signals and lineage that shows when a definition was imported and which downstream answers used it.
For enterprise AI, this release reflects a broader move from generic retrieval towards governed organisational context. The strongest rollout starts with a small set of high-value metrics and questions, measures answer accuracy with subject-matter experts, and expands only after the catalogue proves that it can preserve meaning across source changes and access boundaries.
What teams should validate before production
For implementation, teams should record the exact model, software, dataset, connector and policy versions behind every result. Representative evaluation must include normal work, boundary cases, deliberate misuse and component failure, with outcomes separated by user role, data sensitivity and consequence. A strong average should never hide a serious failure in a critical path. Changes should pass the same regression set before promotion, and operators need a documented fallback that preserves evidence rather than silently broadening access or lowering quality.
Procurement and planning should account for the human work required to review exceptions, maintain controls and investigate incidents. Published performance or efficiency claims are useful starting points, but production confidence comes from evidence collected in the organisation’s own environment. A named owner, measurable acceptance threshold, cost model and tested shutdown path turn a promising release into a service that can be governed over time. Where decisions affect safety, rights, money or regulated records, accountable professional review remains essential.
Security and operational evidence
Security review should map every identity, credential, network destination, stored artefact and administrative access path involved in the workflow. Logs need enough context to reconstruct who initiated an action, which evidence was consulted, what the system proposed and what ultimately executed. Sensitive prompts and traces should not be retained merely because they are useful for debugging. Retention, redaction and access must follow the underlying data classification, with a practical process for correcting or removing material when policy or law requires it.
Operational monitoring should measure quality and service behaviour together. Latency, throughput and availability matter, but so do unsupported claims, missed constraints, unsafe tool calls and the rate at which people overturn recommendations. Teams should rehearse dependency outages and partial failures before unattended operation. If a model, connector or external service becomes unavailable, the workflow should fail predictably, preserve its state and tell the user what remains unfinished. Clear degradation is safer than an agent improvising around a missing control.
Transparency should scale with adoption
Organisations should explain the scope and limitations of the deployed capability to the people affected by it. Users need to know when automation is involved, how to challenge an outcome and where a human can intervene. That transparency improves feedback and prevents a pilot result from being interpreted as blanket approval for broader autonomy. Expansion should follow demonstrated evidence, not novelty or pressure to match a competitor announcement. The durable advantage comes from dependable integration, careful measurement and learning that survives the next model or product update.
Finally, teams should schedule a review after real usage rather than treating launch as the end of governance. The review should compare expected benefits with observed quality, cost, incidents and user behaviour, then retire workflows that do not justify their complexity. A service that works well can expand through the same evidence-based process. A service that fails should be narrowed or stopped without losing the records needed to understand why. This feedback loop is what turns experimentation into accountable operational capability.