Broadcom AAI v26 Adds Conversational Insights and Cost Attribution

Broadcom has released Automation Analytics & Intelligence v26

Broadcom announced the general availability of Automation Analytics & Intelligence (AAI) v26 on August 26, 2026. The release adds a conversational interface over workload data, financial cost attribution, official OpenShift container images and an AI-assisted source-code security process.

AAI is an analytics and control layer for enterprise workload automation rather than a general-purpose chatbot. The most notable AI feature in v26 is designed to make operational data easier to query without replacing deterministic data access.

Generative AI builds the query; deterministic systems retrieve the data

AAI v26 adds natural-language exploration so an operator can ask questions about workload performance and system health without manually constructing specialized queries. Broadcom says generative AI is used only to construct the query, while the actual data retrieval uses direct, deterministic access.

That architecture is significant for operational analytics because it keeps the probabilistic model out of the final data-retrieval path. Teams still need to test whether generated queries correctly represent user intent, but the returned operational records are not generated by the language model itself.

Users can retain conversation histories and publish useful conversational queries as standard data insights, turning exploratory questions into reusable operational views.

A new financial model maps automation execution to business costs

The release also adds explicit cost accounting. Administrators can define variables such as software, infrastructure and labor costs, then use rule-based filters to map job executions to internal cost centers. Chargeback and showback reports can attribute workload expenses to business units.

This moves AAI beyond reliability and SLA analytics toward FinOps-style measurement for enterprise automation. Actual cost accuracy will still depend on an organization's allocation rules, labor assumptions and infrastructure accounting.

OpenShift containers and AI-assisted security are included

Broadcom now provides official container images for AAI components, including an automatically configured single-instance OpenShift setup. Containerized AAI can connect to existing Oracle, Microsoft SQL Server and PostgreSQL databases and to automation engines outside the cluster. Traditional installation and upgrade paths remain supported.

Broadcom also says AAI v26 is its first release to use frontier AI models to scan source code for potential security vulnerabilities and remediate findings before general availability. That is a description of Broadcom's development process, not evidence that AI scanning makes the released software vulnerability-free.

Other additions include secure Airflow proxy support, longer SLA trend analysis and dependency-tracking improvements.

For enterprise automation teams, v26 is notable for combining natural-language operations analytics with deterministic retrieval, financial attribution and modernized deployment. The practical evaluation should focus on query correctness, access control, cost-model quality, OpenShift operational fit and how AI-assisted security findings integrate with conventional testing.