Thirty thousand people will be at Moscone next week, and most of the coverage will be about the keynote demos. For healthcare and life sciences leaders, the demos are not the point. The harder question underneath them is.
Databricks has spent the last few years moving its own story forward: from consolidating data, to querying it in plain language, to letting agents act on it. Ali Ghodsi has a name for that last step, Agentic Data Work, a shift from data professionals being assisted by AI to AI agents doing the work, guided by humans. For a regulated industry, that last step is the hard one. It is one thing to let an agent answer a question about your data. It is another to let it act on patient information, under audit, inside your governance perimeter. With a dedicated HLS track again this year set against an agenda dominated by agents, that is the tension worth watching.
Here are three things worth watching for, regardless of what stack you run.
The featured AI lineup is full of agent companies, and there is an OpenAI-hosted hackathon dedicated to building agentic data apps. That is the easy part to get excited about. The hard part, and the part that matters for a regulated enterprise, is whether an agent can run inside your governance perimeter rather than around it.
We saw the early version of this with Genie Code earlier this year, where deep Unity Catalog integration let an agent understand data semantics, governance rules, and access controls from the inside. Watch whether the agent story at Summit extends that pattern. An agent that reads your patient data from the outside is a demo. An agent that respects consent, lineage, and access controls from the inside is a product you can deploy.
Agents need somewhere to keep track of what they are doing, update it at speed under the same governance. Lakebase, positioned as Postgres for data apps and AI agents, is the piece most HLS teams are not yet paying attention to. If your agent strategy depends on bolting a separate operational database onto your analytics platform, you have just created a second governance surface to audit. Watch how the operational and analytical layers come together, because in life sciences the audit trail is not optional.
Vertex Tower, Plot no-
C-33, 4th Floor, Phase 2, Industrial Area, Sector 62, Noida, Uttar Pradesh, 201309