Storage teams need to know more about their unstructured data, especially when most (74%) of enterprises are storing more than 5PB of unstructured data. Common questions include: how much data do we have, where is it stored, how much is old/cold, how fast is it growing and what is it costing IT to store and protect.
Across NAS, cloud and object silos, that means correlating data across scripts, vendor consoles and spreadsheets. AI tools deliver a faster way to get the answers, but connecting one vendor MCP per storage system loads the model with tool definitions and gives it only a slice of the data estate. That’s MCP bloat and it delivers your teams a suboptimal experience.
Komprise Universal File MCP gives users the ability to use their favorite AI tool (aka ChatGPT, Claude, Perplexity, Gemini, Copilot, Cursor) in one interface to query across all file and object storage. Read the press release.
It runs on the Komprise Global Metadatabase, which already indexes your data estate, so no data moves to answer a question and it is delivered in a consistent schema. Komprise Universal File MCP can deliver smaller, curated results to reduce context bloat and AI tokens, with our built-in metadata extraction capabilities. As well, Komprise Transparent Move Technology keeps tiered and archived data available for query in our MCP solution, with no rehydration.
Example IT prompts with Komprise Universal File MCP
1. Capacity and Footprint
– “How much data do we have, including tiered data?”
– “Show me our storage footprint capacity by shares or sites.”
– “Which shares are using the most storage?”
2. Location
– “Where is our data located?”
– “Show me all the data in Atlanta and where it’s stored.”
– “Where are the image files that were tiered this year?”
3. Data Age and Cold Data
– “How much of our data is more than a year old?”
– “How much of the data is cold, not touched in 2+ years, and where is it?”
– “How much of my data is orphaned?”
– “Find research data by project tag that is over six months old and could be tiered with approval.”
4. Cost and Savings
– “How do we save money?”
– “Give me a strategy to save 50% of my on-premises Everpure, NetApp and Isilon storage space.”
– “Find all videos consuming flash capacity.”
5. Compliance, IP and AI Readiness
– “What can we safely send to AI without violating regulations and internal policies?”
– “Find all IP data, tag it and set up a workflow to search for PII (driver’s license #, email addresses) in that IP.”
– “Find files containing PII, tag them and exclude them from AI.”
6. From Answer to Action
– Set up a Smart Data Workflow to tag all data more than 6 months old and move it to archives
– “Send this curated dataset to our AI pipeline.”
Fold Permissions into Every Prompt
Every query runs under the requesting user’s existing access permissions and Komprise can exclude sensitive data based on the AI use case. IT can also authorize non-IT users to run queries and actions on the data they already have access to see and edit.
Progressive response keeps the context window lean. You can ask Komprise to filter out redundant, obsolete and trivial (ROT) data before it reaches the model.
What to Do Next
Understand the five questions your team answers most often, then try them in your AI tool. Komprise Universal File MCP is available now in an early access program for customers and partners.
Contact your account team or email eap@komprise.com, and learn more at komprise.com/mcp.
