The AI conversation is evolving. Organizations are no longer asking whether they should adopt AI. They are asking how to scale it responsibly, control costs, and accelerate business impact.
At this year's Data + AI Summit (DAIS), Databricks announced several capabilities that directly address these challenges. As Director of AI Engineering at Aimpoint Digital, I left the conference convinced that a few of these announcements have the potential to significantly change how enterprises build, govern, and scale AI solutions.
Below are the three announcements that stood out most to me and why I believe they matter.
1. Unity AI Gateway Spend Controls: Don’t Let AI Costs Hold You Back
One of the most exciting recent announcements from Databricks is the addition of visibility into token cost expenditure. Recent news stories have highlighted organizations incurring unexpected AI costs, including reports of over $500 million in token spend in a single month due to a lack of guardrails, governance, and oversight of AI usage.
Fortunately, Databricks AI Spend Controls through Unity AI Gateway provide the capabilities organizations need to be proactive about managing costs rather than reactive. Organizations can track AI spend at every level of granularity, from individual users to the account level, and across AI use cases ranging from coding agents supporting developers to production AI systems.
You can also configure email notifications for monthly spend thresholds, both at the individual user level and across shared account-level budgets, helping teams monitor usage and take action before costs become an issue.
2. Omnigent: A Unified Harness for AI Development
Omnigent is a one stop shop coding harness that simplifies your AI development workflow. You no longer need to keep your laptop running overnight while your agents execute, as the workloads are hosted and managed directly within Databricks.
The platform also enables more powerful development workflows by allowing you to leverage multiple coding harnesses together. For example, you can use Claude to generate code and Codex to review it, creating a collaborative AI-assisted development experience. Teams can further enhance quality through peer code reviews and comments directly within the workflow.
From a governance perspective, Omnigent provides additional controls by allowing token usage limits to be configured at the session level, helping organizations better manage AI spend and operational costs.
This is one of the announcements I am particularly excited to explore further, and I am looking forward to seeing it in action within my own Databricks applications.
3. Genie Ontology: A Trusted Context to Enterprise AI Agents
Your agents are only as good as the context they receive. Providing agents with a consistent, unified view of organizational knowledge remains one of the biggest challenges in enterprise AI. Today, this process is often time-consuming and unreliable due to fragmented data sources and the non-deterministic nature of agentic systems.
Genie Ontology was one of the most exciting announcements aimed at addressing this challenge. It builds a knowledge graph that connects all asset types across an organization, including external applications such as Slack, Microsoft Teams, and more. Under the hood, it is powered by Databricks' OntoRank algorithm, which is conceptually similar to Google's PageRank algorithm but designed specifically for enterprise data and assets rather than webpages.
OntoRank generates an authority score based on factors such as asset ownership, utilization, and relationships to certified assets. Databricks reported that, internally, Genie Ontology delivered more than a 30% improvement in query accuracy. Beyond improving accuracy, graph-based traversal enables more targeted context retrieval, which can reduce token consumption and lower overall inference costs.
Genie Ontology powers Genie One, Databricks' AI coworker, and is deeply integrated with Unity Catalog for governance through lineage, semantic glossaries, and domains. By grounding AI systems in trusted organizational knowledge, Genie Ontology helps agents retrieve the right context at the right time.
Be on the lookout for more Aimpoint Digital blog posts covering Genie One and Genie Ontology. If you want to learn more, reach out to our experts!
.png)
