Meet Preset Chatbot: Your Enterprise-Ready Data Conversation Partner
When we announced Preset MCP a few weeks ago, we talked about AI that doesn't just answer questions — it builds. Today, we're turning that idea into a generally available product. Preset Chatbot is here and ready for enterprise prime time.
This is a big moment for us, and a genuinely big moment for how your teams work with data.
Ask your data a question. Get an answer.
Preset Chatbot is a conversational analytics assistant built directly into Preset. You ask a question in plain English, and it answers with a chart, a dashboard, or a written insight backed by your actual data.
That's the whole pitch, and using it really is that simple:
- Ask questions, get charts. "What were our top-performing campaigns last quarter?" turns into an interactive visualization you can explore, filter, and refine.
- Build dashboards through conversation. Start with a question, add another, iterate on what you see, and end up with a full dashboard without ever leaving the chat.
- Generate SQL on demand. If you need to drop into the SQL IDE, Chatbot writes the query for you. Copy it, tweak it, ship it.
- Explore datasets without writing SQL. Chatbot understands your semantic layer. It knows which columns mean what, which metrics are trustworthy, and how tables relate.
This isn't a generic LLM bolted onto a BI tool. It's an analytics-native assistant grounded in your data model and your governance rules. (See more of what Preset Chatbot can do →)
Why this matters
For years, self-service BI has been the dream and, honestly, the lie. Vendors have been promising it since Excel got pivot tables. And yet in most organizations, answering a question like "how did our cohort retention change last month?" still means opening a ticket, waiting for a data team to run a query, and receiving a screenshot three days later.
That friction costs organizations in two directions at once:
- Business users can't move at the speed of their own curiosity. They either get stale answers or go without.
- Data teams spend their time on repeat ad-hoc requests instead of on the work that actually compounds: better models, richer semantics, more useful metrics, better tooling.
Preset Chatbot is a real answer to both. Business users get instant, self-service access to their data. No SQL, no ticket queue, no waiting for someone else's calendar to clear. Data teams get their focus back.
This isn't about replacing data teams. It's about freeing them. When the chatbot handles the "show me weekly active users broken out by plan" questions, your analysts can work on what actually makes your data platform better: the modeling, the semantics, the governance, the metrics that matter. Chatbot is a force multiplier, not a replacement.
We see this as another leap in democratizing access to data. Dashboards put BI in the hands of anyone who can click. Chatbot puts it in the hands of anyone who can ask a question.
Built on MCP. Built on Superset. Built for enterprises.
Under the hood, Preset Chatbot is powered by Preset MCP, our implementation of the Model Context Protocol that gives AI safe, governed access to the full analytics platform. If you've been reading along with our MCP series, you already know the punchline: we wrap, we don't fork, and we inherit all the security and RBAC machinery that makes Preset enterprise-ready.
That means Chatbot respects:
- Row-level security. Users only see data they're allowed to see, even in a chatbot conversation.
- RBAC and workspace permissions. The same access controls as the rest of Preset.
- SOC 2 and audit logging. Whatever Preset already audits for human users gets the same treatment when Chatbot is the actor.
- Your semantic layer. Answers are grounded in the datasets and metrics your data team has blessed, not arbitrary SQL against raw tables.
No hallucinated columns. No phantom metrics. No surprise data exfiltration. It's AI that respects the guardrails your data team already put in place.
And because Preset is built on Apache Superset®, Chatbot inherits that same open-source foundation: visualizations, SQL IDE, semantic layer, the works. (There's a real reason the top AI models know Superset so well. See Max's piece on open source's massive unfair advantage in the AI era for the long version.)
If you're curious about the deeper "how," we've written a few technical companion pieces:
- Preset MCP: From Open Source to Enterprise: how we wrapped Superset's MCP service for multi-tenant, enterprise use.
- Superset MCP Service Deep Dive: the OSS-level technical reference.
- Building Better BI Chatbots: what actually makes a BI chatbot useful (vs. just demoable).
How to turn it on
Preset Chatbot is a workspace-level feature you enable in Preset Manager.
Head to Preset Manager → Workspace Settings on the workspace where you want Chatbot enabled. You'll find a tab there to flip the feature on.
Once it's on, the chat panel shows up in your workspace and your users can start asking questions.
Availability and rollout
We're launching Preset Chatbot on our Enterprise plan first, and we'll expand access to more plans over time. Here's what that looks like in practice:
- Existing Enterprise customers: flip Chatbot on yourself in Workspace Settings to try it out, then reach out to your Customer Success representative when you're ready to upgrade to full access.
- New Enterprise trials: you'll get a limited pool of tokens allocated automatically when your trial starts. Flip the feature on in Workspace Settings, and you're off to the races.
- Pro and Trial customers: your workspace admin can turn on a limited evaluation of Chatbot from the same Workspace Settings tab, so your team can kick the tires before committing to Enterprise.
- When your trial ends: Chatbot functionality is disabled. If you need additional tokens to keep exploring, or if you'd like to extend your trial, contact Preset Customer Support. We're happy to help.
We'll share more on broader availability in the coming months. For now, Enterprise customers get first dibs.
Frequently asked questions
Who can use Preset Chatbot today, and when will it reach other plans?
Preset Chatbot is launching on our Enterprise plan. Existing Enterprise customers can flip it on themselves in Workspace Settings to try it out, then reach out to their Customer Success representative when they're ready to upgrade to full access. New Enterprise trials will have tokens provisioned automatically. Broader plan availability is on the roadmap. We don't have specific dates to share yet, but Enterprise is just the starting point.
What happens when my trial ends or I run out of tokens?
When your Enterprise trial ends, Chatbot functionality is disabled on the workspace. If you use up your allocated tokens during the trial, or want to extend the trial to keep exploring, reach out to Preset Customer Support. We can usually make it happen.
Which data warehouses does it work with?
All the ones Preset already supports: Snowflake, BigQuery, Redshift, Databricks, Postgres, and the rest of the Superset connector catalog. Chatbot sits on top of your existing semantic layer, so if Preset can query a source, Chatbot can answer questions about it.
Can I use Claude, Cursor, or other AI tools alongside Chatbot?
Yes. Preset Chatbot is the in-product experience, but Preset MCP exposes the same governed analytics platform to any MCP-compatible AI client. Use whichever AI surface fits your team's workflow. Or use both.
Can I bring my own model provider?
Yes. Paying Enterprise customers can plug in their own OpenRouter API key to route Chatbot traffic through the account of their choice and handle token billing on their side. BYOK isn't available on Free, Trial, or Pro tiers at GA.
See it in action
If you'd rather watch us drive before you take the wheel, we're hosting a live demo on May 6.
The bigger picture
This is step one, not step ten. Preset Chatbot is the first user-facing surface of a broader bet we're making: that the future of analytics is conversational, agentic, and grounded in the same governed semantic layer that already powers your dashboards. Chatbot is where that bet shows up for business users today. MCP is the foundation that lets Chatbot (and whatever comes next) plug into the AI tools your teams already live in.
Preset is all-in on what AI brings to analytics workflows, in our ongoing effort to democratize access to data and the insights it can bring to your entire org. We're leveraging Chatbot every day, and are even building our own agents using Preset MCP. We've been working toward this moment for a long time, and we couldn't be more excited to put it in your hands. We can't wait to hear your feedback, and show you what comes next.
