Channel partners deploying AI agents need to connect customers’ business data across systems while managing integration costs, access permissions, and data quality.
Google Cloud announced on Nov. 4 that it is expanding its Agentic Data Cloud with updates to its borderless Lakehouse and Knowledge Catalog. The changes let enterprises ground AI applications in supported data sources without migrating the underlying datasets, although retrieved data blocks can be cached in Google Cloud.
Google also announced AlloyDB integrations that let developers query live operational data alongside historical Iceberg data. Separately, AlloyDB Express entered preview, offering developers and AI coding agents a PostgreSQL database that automatically pauses when idle and can be upgraded to AlloyDB as application requirements grow.
“By implementing the borderless Lakehouse that federates single-copy Apache Iceberg data from Amazon S3 tables to BigQuery, Delivery Hero’s talabat reduced data-movement costs by approximately 40% while making streaming transactional events queryable across both clouds within minutes of production,” said Harish Ramesh, staff data engineer at Delivery Hero’s talabat, in Google’s announcement.
Google’s Agentic Data Cloud gets new capabilities
The new Agentic Data Cloud capabilities connect distributed enterprise data, ground AI agents in business context and support agent workloads at scale.
Enterprise data sits across multiple clouds, SaaS applications, and on-premises environments. The borderless Lakehouse supports querying data stored in AWS and Microsoft Azure, alongside catalog federation and integrations with supported SaaS platforms.
To derive the most value from AI operations, agents need business context to produce accurate answers and take appropriate actions. Google Cloud says that it has extended its Knowledge Catalog with “self-learning, reusable context sets, and a reimagined user experience.” The cloud provider further says that Knowledge Catalog makes AI agents nearly 4x more accurate, according to its internal assessment.
Cross-cloud interconnect and caching for AWS and Azure are now generally available. Knowledge Catalog is also on by default in Gemini in preview.
What channel partners should validate and package as services
Partners should map each customer’s data sources to the supported integrations, catalogs, and networking options before recommending a deployment. They should also identify which required capabilities are generally available and which remain in preview, then confirm that their support commitments match the customer’s production requirements.
Cost assessments should cover query processing, networking, caching, catalog usage, and the staff time needed to operate the environment. Partners reviewing infrastructure costs should distinguish hardware budget pressures from the cloud consumption charges that determine an agent deployment’s ongoing cost. Keeping source data in place does not remove the need to model those charges.
Data access also needs a business-level check. Partners should work with customer data owners to agree on definitions for measures such as revenue or customer churn, validate the records agents will use, and test whether permissions prevent access to restricted information. A technically successful connection can still produce an unusable answer if the underlying definitions or data are wrong.
Before expanding deployment, partners should pilot one workflow and measure answer quality, response time, and operating cost against the customer’s existing process. Agreeing on acceptance criteria upfront gives both parties a clearer basis for deciding whether the deployment delivers enough value to continue.
These requirements could support assessment, integration, governance, and ongoing optimization services. To make those services commercially viable, partners should define who maintains business definitions, reviews access changes, investigates incorrect answers, and monitors spending. Clear ownership and service boundaries help customers understand what ongoing support covers and help partners price the work realistically.
Read more: Turning AI deployments into lasting customer value requires measurable results. Explore how AWS’s AI ROI push connects partner incentives with ongoing governance and performance reporting.