MetaSpark Brain: Task Ranking Without the Lies

The Integration Trap

Every product team ships integrations. Asana has them. Jira has them. Linear has them. But integrations solve a shallow problem: they move data from one system to another. They don't answer the question every founder, engineer lead, and ops manager asks every morning: what should I actually work on first?

The integration catalog means nothing if you're drowning. More systems connected just means more signals to ignore. Your Linear board has 40 open tickets. GitHub has 30 PRs. Notion has an unread deal summary. Slack has three threads you need to respond to. Your calendar has a 10-minute hole at 2 PM. What gets done today?

That's not a data-movement problem. That's a prioritization problem. And prioritization has been broken in every work-tracking tool that ever existed.

How MetaSpark Brain Works

MetaSpark v2 introduced the Brain as the ranking layer underneath our entire platform. Brain scores every task in your system, across all connected boards and tools, against three competing signal families.

First: calendar fit. Brain reads your calendar and asks: do you have a contiguous focus window today that's large enough to finish this task? A 15-minute code review is doable in the gap between standup and a 1-1. A design spec isn't. Brain knows the difference. If you don't have calendar space today, the task stays in the queue, no matter what else is true about it.

Second: dependency graph. Brain builds a directed graph of what blocks what. If a design review is waiting on a stakeholder's sign-off due in three days, but that sign-off unblocks a PM and two engineers, it ranks high. If your task is waiting on someone else who's on PTO, it ranks low. If your task is blocking something on a near deadline, it rises. Dependencies aren't entered manually, Brain infers them from the commit history, the PR threads, the comment patterns in Linear and Notion.

Third: completion patterns. Brain watches what you actually finish. Not what you say is urgent. What you ship. If you consistently complete code reviews within a day but design specs take a week, a code review that arrived yesterday is ranked higher than a spec that arrived last week, all else equal. This signal is ruthlessly empirical and nearly impossible to game.

The three signals run every time the calendar shifts, a dependency resolves, or a new task lands. The ranking re-stabilizes in near-real-time.

Why This Matters at Scale

At a 50-person engineering team running Linear, GitHub, Slack, Notion, Gmail, and a custom internal CRM, a prioritization field in any single tool is theater. The same task lives in four places. It gets marked "urgent" by four different people in four different urgency schemes. You check the wrong tool first.

Brain solves this by making your ranked list the single source of truth. Every task in the system, no matter where it was born, gets one score and one position. The same task doesn't appear four times with four different urgencies; it appears once, ranked against everything else you could be doing.

Ops teams see this immediately. A customer escalation in Zendesk is blocking two engineers and due in four hours; a half-day debugging session you've been planning for two weeks is waiting on a Figma file that's still in review. Which gets the team's focus today? Brain knows. The Zendesk ticket rises. The debugging session waits.

Engineering leads see it too. A PR in GitHub has been sitting for three days, is blocking a junior engineer's work, and you have a 90-minute block this afternoon. Brain surfaces it. An internal metrics dashboard task has been waiting for a week, but you're booked solid this week and no one else needs it before Thursday. Brain keeps it below the fold.

Brain Inside the MetaSpark Harness

Brain doesn't live in isolation. It's the ranking engine that powers the entire MetaSpark platform. Our agents use Brain's scores to decide what to work on next, which PR to review, which issue to pick up, which customer thread to draft a response for. When you configure auto-accept thresholds for agent execution, you're often saying: "agents can handle this class of task if it's in the top 20% of what I should be doing today." Brain's score is what determines that.

Multi-Board Views queries the same ranked graph: your "My Day" board shows the top 8 items Brain has ranked for you. An Engineering lead's view shows only engineering-class tasks ranked the same way. A Customer Ops manager's view filters to customer-facing work. Same ranking engine, different lenses.

The audit trail in Live Task Triage shows every rerank. When a task moves from position 12 to position 3, you see why: a dependency resolved, your calendar opened up, or someone added a comment signaling urgency that Brain's pattern-matching picked up.

What Changes

Before Brain, you lived in Slack asking your team: "What should I focus on?" or in Linear re-ordering your own list based on gut feel. The ranking was manual, inconsistent, and the list you left at 5 PM wasn't the same list you looked at when you opened your laptop the next morning.

With Brain, the list re-ranks every time the world shifts. You open MetaSpark and the top item is what you should actually work on right now, not what felt urgent yesterday.

For engineering teams, this means fewer planning meetings. For ops teams, it means faster escalation response. For founders, it means the company's most critical path gets focus first, automatically, every single day.

Try It

Brain ships today as part of MetaSpark v2. Log in to your MetaSpark workspace, connect your first board (Linear, GitHub, or Notion), plug in your Google Calendar, and watch the ranking stabilize. The audit trail in Live Task Triage shows the signals moving in real-time.

Want to bring your own agent? Brain's rankings are available over the MetaSpark API and Model Context Protocol. Build your own planner; let Brain's scores drive your routing.