Preset MCP: What 25 Tools Actually Means for AI-Powered Analytics
Every BI vendor in 2026 has an AI story. Most of them sound the same: ask questions in natural language, get charts. It's a fine start. But buyers should be asking a harder question: what can your AI actually do?
It's a bigger distinction than it sounds. An AI that can only read your data is a fancy search box. One that can write back to it is actually doing the work alongside you.
MCP: The Protocol That Makes AI Composable
Model Context Protocol (MCP) is an open standard that lets AI clients — Claude, ChatGPT, Cursor, or any custom agent — connect to external tools in a structured, secure way. It's a universal adapter between AI models and the systems they need to work with.
For BI, MCP means an AI agent can do more than translate your question into SQL. It can discover what datasets exist, understand their schema, query them using governed metrics, and create new data assets that persist after the conversation ends. We've written before about the promise of MCP-powered data workflows; this post is about what changes once those tools can write, not just read.
In April, we shipped Preset's MCP server with 20 tools. In May, we expanded to 25, organized across five domains: system health, discovery, read operations, chart and dashboard management, and dataset/SQL operations.
Those last two, the dataset and SQL operations, are the ones worth talking about, because they're the ones nobody else ships.
Most BI MCP Servers Are Read-Only. Ours Isn't.
We surveyed every major BI vendor's MCP implementation: Tableau (16 tools, GA), Power BI (20+ tool categories, Public Preview), Looker (managed, Preview), ThoughtSpot (GA with Spotter 3), Sigma (GA), Metabase (GA), Hex (Beta), Omni, and Lightdash. I went through all of them, and most amount to a handful of read-only tools and a press release. Every one focuses primarily on querying existing content: find a dashboard, run a report, summarize a chart.
Preset's MCP server does that too. It also does something none of them do: it creates.
create_virtual_dataset
This tool lets an AI agent save a SQL query as a reusable virtual dataset in Preset. Not a one-off query result, but a governed, named dataset that other users, dashboards, and future AI sessions can build on.
Most AI analytics workflows today are disposable. You ask a question, get an answer, and the work evaporates. The next person who needs the same insight starts from scratch.
create_virtual_dataset turns AI conversations into durable data assets. An analyst working in Claude or Cursor can explore a question, refine the SQL, and then save the result as a first-class dataset in Preset, complete with column descriptions, metrics, and governance. That dataset is immediately available to every other user and every future AI session.
No other BI vendor's MCP server offers this capability today. Tableau's MCP reads datasources and workbooks. ThoughtSpot's creates analysis sessions for conversational reasoning. Looker's queries governed models. All useful. All read-only. Preset is the only one where the AI can actually add to your data layer.
query_dataset
This tool lets an AI agent query datasets using saved metrics, calculated columns, and dimensions instead of raw SQL. The AI works through the semantic layer, respecting the definitions, calculations, and business logic your team has already established.
This is the difference between "run this SQL" and "query revenue by region using the metric definitions my team agreed on." The latter is what you need if you're going to trust the number that comes back.
get_chart_sql
Transparency is a governance requirement. This tool lets an AI agent retrieve the underlying SQL and applied filters for any chart. When a stakeholder asks how a number was calculated, the AI can show them the exact query, the exact filters, and the exact data source. Most competitors don't expose this level of transparency through their MCP servers.
What This Looks Like in Practice
Here's a workflow that's possible today with Preset's MCP server and Claude:
- A data analyst opens Claude and connects to Preset via MCP.
- They ask: "What tables do we have related to customer churn?"
- Claude uses Preset's discovery tools to find relevant datasets, returning schema and metric definitions.
- The analyst refines: "Build me a query that shows monthly churn rate by customer segment for the last 12 months, using our standard churn metric."
- Claude uses
query_datasetto run the query through the semantic layer, respecting the team's metric definitions. - The analyst reviews the results, adjusts the segmentation, and says: "Save this as a virtual dataset called 'Monthly Churn by Segment' so the customer success team can use it."
- Claude uses
create_virtual_datasetto save it: a governed dataset, visible in Preset, queryable by anyone with the right permissions.
That last step is the one no other BI vendor supports through MCP. The dataset doesn't disappear when the conversation ends. It becomes part of the organization's data layer.
Open Source Foundation, Enterprise Execution
Preset's MCP server is built on SIP-187, an open-source specification contributed to Apache Superset. Any MCP client that works with open-source Superset works with Preset, and vice versa.
This is the part competitors can't easily copy. Tableau's MCP works within the Salesforce ecosystem. Looker's routes through Google Cloud. Power BI's lives in the Microsoft Fabric universe. Preset's MCP works with any AI client, on any cloud, connected to any of our 75+ supported databases.
For organizations running multi-cloud environments or evaluating AI tools across their stack, that neutrality is a requirement. Your BI tool's AI capabilities shouldn't lock you into a specific model vendor or cloud provider.
Security by Design
Every MCP tool call goes through the same security stack as the Preset application itself: OAuth 2.0 with PKCE, role-based access control, row-level security, and full audit logging. An AI agent connected via MCP can only access the data its authenticated user is authorized to see.
The MCP server inherits Preset's seven-layer middleware architecture — the same governance that supports SOC 2 compliance, HIPAA-eligible deployments, and enterprise-grade access control. When an AI agent creates a virtual dataset, that dataset inherits the workspace's security model automatically.
The Bigger Picture
The BI market is converging on a single question: what role does your analytics platform play in an AI-native workflow?
Some vendors are positioning as the destination, where AI features live inside the product UI. Others are positioning as the source, a governed data layer that any AI tool can query. It's a question we explored in AI in BI: the path to full self-driving analytics.
We think the answer is both. Preset Chatbot gives teams an in-product conversational experience grounded in the semantic layer. Preset MCP makes the platform composable: an open building block that AI agents, developer tools, and custom applications can connect to and build on.
Twenty-five tools that read, query, discover, and create, on an open protocol with an open-source foundation and enterprise security underneath. That's a longer answer than "we added a chatbot," which is rather the point.
Want to see it in action? Read the Apache Superset MCP technical deep dive, see the full Preset MCP announcement, or go under the hood in the Preset Chatbot technical deep dive.
