Video
Meet the AI Assistant in Perforce Delphix Data Control Tower [DEMO]
Data Management,
Security & Compliance,
AI
Back to top
Overview: Getting to Know the Delphix AI Assistant in Data Control Tower
Managing and refreshing test data environments shouldn’t require memorizing APIs or filing tickets. The integrated AI Assistant in the Perforce Delphix Data Control Tower (DCT) 2026.6, makes it simple for teams to manage their data environments in plain language.
In this demo, Principal Sales Engineer José Rodrigues shows how administrators configure the Delphix AI Assistant and how end users put it to work. Learn how you can enable the AI assistant, assign LLM providers to individual AI features, use multiple providers and models for different features, and refresh a virtual database with tracking from start to finish.
With this AI assistant, you can get your teams get the data they need, faster.
Take Advantage of the Delphix AI Assistant & Capabilities
For organizations pursuing agentic development while balancing compliance, Delphix will help secure your data and enable new ways to use AI agents. See how our DevOps Data Platform and solutions make it easy for your teams to get trusted, compliant, and AI-ready data.
Back to top
Full Transcript
Hello everybody and welcome. My name is Jose Rodriguez and I am a Principal Sales Engineer here at Delphix, Perforce.
Today I am going to run a quick and short demo of the new feature released with DCT 2026.6. This is the integrated Data Control Tower AI Assistant. Do not confuse this with the DCT MCP server we also have out there.
In the demo I will be showing how the assistant is configured and how it is accessed. Note that the assistant is context-aware, meaning that it may behave differently depending on what part of the UI we find ourselves in.
Also remember that this is AI, and AI makes mistakes. Always double-check the response before accepting a potentially destructive action.
Let’s review now the AI settings that enable this feature.
In this demo version of DCT, you can see that different features in DCT can have their own LLM model assigned. I am showing this with an administrator user, which has the power necessary to configure everything.
So you can see that we can enable the new assistant here, Delphix AI Assistant, but also other pieces of assistants and AI that are available in DCT.
I have only one provider configured here, but you can add as many providers as you see fit, potentially using different models for different features.
Now that we have the AI settings configured, let’s go and see how a regular user can leverage the new feature.
I have now started a new DCT session with an account that has limited access to only self-service their own VDBs.
I will quickly demonstrate how easy it has become for an end user to refresh their own VDBs. I will first be asking for a list of VDBs and whether they are masked.
For each of these prompts, we have the option to show the details of the different API calls the system is actually making.
This is interesting for auditing and potentially automating this process later on.
I am now reviewing the list of snapshots available for me to refresh the VDB from its parent.
Next, I refresh the VDB without indicating any snapshot.
This is to demonstrate that the assistant does not blindly move forward, as there are guard rails in place.
Here we can see that for any potentially destructive operation, the AI assistant will always require human validation.
We now have a job ID in DCT, which we can track with the AI assistant.
At this point the job has just started, 43% complete.
Started by me in this particular engine and with this snapshot.
We can check again, and ask for a job summary.
Status: completed, 100%, roughly 25 seconds to complete, and other details that may be of interest.
This is the end of the demo. Thank you for your time today.