Using AI to Help Manage your Data Infrastructure (Ask Us Everything Recap)
AI is showing up everywhere in IT. But for storage admins, the real question is whether it can actually help save time and make managing infrastructure easier.
That was the focus of our latest Ask Us Everything, where Everpure experts Anand Madhavan (Director of Product Management), Ameerul Shah (Product Manager) and Russell Cagle (Tech Lead), fielded questions from the Everpure Community about putting AI to work in real-world storage environments.
What can Pure1 AI Copilot actually do?
The discussion started with the basics: How was this tool designed to help IT professionals?
Anand described it as the next step in Everpure’s long-running focus on simplicity. Instead of searching documentation, digging through multiple screens, or remembering commands, IT folks can ask questions in natural language. Pure1 AI Copilot uses information from Everpure knowledge sources along with data about your own fleet to help find the answers you need.
That means you can ask important things like: What’s the latency on this array? Which arrays need an upgrade and why? ? What alerts should I be concerned about? No deep knowledge is needed to get actionable answers.
Importantly, Pure1 AI Copilot is currently focused on finding answers, and giving you options. It can reason out a problem and recommend what to do next, but it doesn’t simply start executing commands. And this is by design: Everpure’s approach separates AI reasoning from deterministic execution, so there aren’t any surprises. Pure1 AI Copilot works with capabilities such as Workflow Orchestration and Everpure FusionTM Presets to automate tasks, so you stay in control.
Can AI really make troubleshooting faster?
Consider that familiar question: “My array was slow last night. What happened?”
Traditionally, answering it might require piecing together telemetry, documentation, CLI information, alerts, and possibly opening a support case. Triage Assistant is designed to work through that investigation process for you—checking alerts and upgrades, looking for latency spikes, identifying busy volumes, and then providing evidence and recommended next steps.
Ameerul shared that an investigation that might otherwise stretch across a day or two could potentially produce useful answers through Copilot in a matter of minutes.
That’s the kind of AI benefit storage teams can appreciate: less hunting for information and more time spent on solving problems.
What if I want to use my own AI assistant?
This is where the conversation turned to MCP servers.
Russell explained that the Pure1 Manage MCP server gives compatible AI assistants (like Claude, Cursor, etc.) access to Pure1 Manage data. That opens some intriguing possibilities because storage rarely operates in isolation.
When asked whether Pure1 AI Copilot could understand environments beyond Everpure, such as VMware and Azure, the answer highlighted the bigger opportunity behind MCP. That is, connecting Everpure data with MCP servers for other parts of your infrastructure, so an AI assistant can correlate information across storage, virtualization, networking, observability, and other systems.
Instead of becoming yet another management silo, data on Everpure storage can become part of the AI tools that IT admins are already using.
Do I need an MCP server for every array?
For admins managing larger environments, one attendee had a very practical question: If I have 18 arrays, do I need 18 MCP servers?
No—and that answer gets to an important difference in Everpure’s approach.
With Everpure Fusion, the focus is the fleet rather than individual arrays, so its MCP server can work across a fleet containing multiple FlashArray and FlashBlade systems. Configure it once rather than treating every array as a separate management island. And that extends into monitoring fleets—rather than logging into dozens of systems and assembling the picture yourself, Pure1 AI Copilot can help provide a fleet-wide health view and then let you drill into systems requiring attention.
What’s next?
The panel of experts discussed everything from future support for OpenStack and vCenter correlation, to extending into replication analysis and write capabilities
The takeaway was simple: AI isn’t being positioned as a replacement for the storage admin. Think of it as a “high-performing teammate”: It can help find you information, investigate problems, explain unfamiliar environments, accelerate onboarding, and bring fleet-wide storage intelligence into the AI tools you already use.
For IT teams continually being asked to “get more done with less”, that may be the most useful application of AI of all: spending less time managing the complexity of storage and more time managing the data that matters.
Check out this and all our other Ask Us Everything sessions. And, keep the conversation going by jumping into the Everpure Community.
