Introducing Tasks: your operations, on autopilot

Today we’re launching Tasks in research preview. It’s AI that looks through your company’s data and figures out the specific, actionable things you should do to improve your business.

Dashboards are good at telling you what happened but bad at telling you what to do about it. Tasks reads your revenue, churn, activation, and pipeline, and hands you a prioritized backlog of concrete work. Each task includes the evidence for why it matters now, the outcome you should expect, and step-by-step instructions to get it done.

From “what happened” to “what to do”

Many teams we talk to have the same setup: dashboards full of metrics, a weekly meeting where someone squints at them, and a backlog assembled from whoever spoke last. The data usually shows what’s wrong: 214 trials stalled before connecting a data source, 12 invoices worth $86k are overdue, onboarding is leaking at the invite step. But turning that into a plan is manual and slow, and it’s easy to skip.

Tasks does that step automatically. Click Generate tasks (or let Basedash keep your backlog topped up on its own) and the AI explores your connected data sources, looks at how you already measure the business through your dashboards, and writes a small batch of high-impact tasks, grouped by priority.

Each task comes with three things:

1. Why now: the evidence from your data for why this is worth doing today.
2. Expected outcome: the metric it should move, and by roughly how much.
3. Instructions: concrete steps to execute, written for whoever picks it up.

Send it where work happens

A recommendation is only useful if it turns into work. Every task is one click from your workflow:

• Copy it into your issue tracker. Each task copies as a clean, structured brief (title, priority, why-now, expected outcome, instructions), ready to paste into Linear or wherever your team plans work.
• Hand it to an agent. The same brief is written to be executable, so you can paste it straight into Claude, ChatGPT, Cursor, or whatever agent you use.
• Kick it off with the Basedash agent. For data work, like building the winback segment, drafting the dashboard, or setting up the alert, the Basedash agent can start on it directly.

The loop: Basedash tracks what every task changes

When you mark a task done, Basedash keeps watching the metrics that task was supposed to move. For example, it checks whether trial → paid conversion changed after the winback campaign shipped, and by how much.

Those results feed back into the system. Work that moved the numbers gets doubled down on, so the next batch might suggest scaling the winback play to churned annual accounts. Work that didn’t gets deprioritized. The longer you run the loop, the sharper the recommendations get, because they’re trained on what worked in your business rather than what works on average.

The long-term goal for Tasks is to automate more of your company’s operations, driven by your own data.

About the research preview

Tasks is shipping as a research preview. That means:

• The core loop is ready to use today: generating prioritized, evidence-backed tasks from your data, executing them, and keeping your backlog refilled.
• The impact-tracking and learning side is early and actively evolving. Expect it to get noticeably better over the coming months, and expect some rough edges along the way.
• Your feedback directly shapes where it goes. If a task is off, dismiss it (that signal matters) and tell us what you’d want instead.

Getting started

1. Sign up for Basedash (or log in)
2. Connect your data sources
3. Enable Tasks in your organization settings, then open the Tasks page from the sidebar
4. Click Generate tasks, and add context in settings if you want to steer what kinds of tasks the AI proposes

You can also set an automatic refill threshold, so whenever your pending backlog drops below it, Basedash generates the next batch for you.

What’s next

Tasks joins AI chat for ad-hoc questions, Insights for findings the AI surfaces on its own, and Automations for recurring analyses. Tasks sits on top of them. It takes everything Basedash knows about your business, turns it into what you should do next, and then checks its own work against your metrics.

Try Tasks today to get a to-do list built from your data.