Basedash: AI data analyst
AI-native Business Intelligence Platform
Start new thread

Basedash Actions - A BI tool that can take action for you

byβ€’
Basedash answers questions about your data. Now it acts on them. Ask the agent to extend a trial, fix a record, or seed a demo org β€” it writes the SQL and runs it against any database an admin has enabled for edits. Ask it to update a Stripe subscription or create a HubSpot lead and it acts through any MCP tool you've connected. Every consequential action pauses for your approval, and every tool has its own permission. Skills chain it all into workflows. From answers to actions.

Add a comment

Replies

Best
Maker
πŸ“Œ
Hey everyone, Max here from Basedash. Today we're launching Actions: the Basedash agent can now change things, not just report on them. Turn on Allow edits for a database connection and the agent writes and runs SQL against it: extend a trial, fix a bad record, update the state of a hundred items, spin up a demo org. Connect an MCP server and it takes action in your other tools too: update a subscription in Stripe, create a lead in HubSpot, send an email through Resend. The part that makes this safe to actually use: nothing consequential runs without you. The agent shows you the exact SQL or tool payload and waits for approval, and every MCP tool has its own permission (always allow, needs approval, or blocked) so you decide which actions run automatically and which ones pause for a human. We run Basedash on this internally. Extending a customer's trial used to mean a database edit, a Stripe change, and a follow-up email across three tabs; it's now one skill the agent runs, with one approval. PH community gets an extra week on their trial this week. Happy to answer anything.

Β S/O for this new launch! keep up the great work, and keep launching

Maker

Thank you ! Means a lot coming from you πŸ™

Love that the natural language input sits front and center instead of burying it behind a SQL editor or settings panel. The "describe what you want" framing makes it feel like a creative tool rather than another BI dashboard.

Maker

Β yes exactly, that was super important to us. We market ourselves as AI-native because we really designed and built this product around AI from the start. In many ways, we're much closer to Lovable or v0 than Tableau or Looker.

Maker

Kris from here! Super excited about this launch. We built this because we realized that we often wanted to act on all the insights we were getting from our own product. But carrying out these write tasks is still pretty tedious for a non-technical user, whereas technical teams need guardrails in place before non-coders like me can update the database. So we built both :)

Now the agent writes the fix, shows you exactly what it's about to run, and waits for your approval. One click. All with tons of controls. Admins enable edits per connection. Every tool gets its own permission level. Routine operations run automatically, sensitive ones pause for a human, dangerous ones stay blocked. Your rules, enforced every time.

We've run this internally for weeks and it has already changed how our own team operates. The person closest to the customer fixes the issue.

Happy to answer anything!

An LLM writing and running mutating SQL against a real database is exactly the part I'd want the guardrails on. When it says 'update the state of a hundred items', does it show me the statement and the affected-row count before it runs, and does it wrap the write in a transaction I can roll back? The classic failure is a dropped WHERE turning a 100-row update into 100k, and the model sounds equally confident either way. Preview plus row-count plus rollback is what would get me to flip 'allow edits' on in prod.

Maker

Β yes exactly, you can preview the exact query and see what would be affected before deciding whether you want to approve or reject the action. Reliability and trust are super important for a feature like this.

Preview plus approve/reject covers the intent check nicely. The one that's bitten me: the affected-row count you show at preview is a separate query from the write, so under any concurrent traffic the number I approved and the number that actually changes can drift. Do you run the count and the update inside one transaction, or is that preview count more of an estimate?

The natural language to chart flow feels really polished, like the team actually thought through what happens when a query doesn't quite match the data. That's rare in AI BI tools.

Maker

Thanks , those kinds of details are very important to us!

How does it handle really messy or unstructured data sources, like pulling from a NoSQL database or a third-party API that doesn't return clean tables?

Maker

Β we have an ETL pipeline built in that syncs third-party API data into structured tables. It can also work well with NoSQL databases like MongoDB.

Our AI harness is specifically built to handle these kinds of messy datasets well. It learns the shape of the data over time and improves its own context.

Connected my Postgres database and asked it to show weekly churn by plan in plain English, and it built the chart in seconds. Way easier than dragging fields around in my usual BI tool.

The conversational UI for generating charts feels really considered, you can tell the team obsessed over the small stuff like how follow-up questions modify the existing visualization instead of starting from scratch.

Connecting my Postgres was painless and the natural language chart builder nailed the visualization on the first try. Wish the dashboard sharing flow was a bit smoother though.

12
Next
Last