Setting Your OKRs: How to Give Your AI Workforce Strategic Direction

Six · 5 min read · April 10, 2026

Setting Your OKRs: How to Give Your AI Workforce Strategic Direction

Your Supanova workspace is live. Your agents are hired. But before anything executes, the system needs to know where you're going.

That's what OKRs — Objectives and Key Results — do in Supanova. They sit at the very top of the execution hierarchy. Every project your AI workforce runs, every task it generates, every agent it assigns traces back to an objective you set or approved.

This guide walks you through setting your first OKRs, how the system uses them, and why the structure matters for building an AI workforce that gets smarter the more you use it.

The Hierarchy: Why OKRs Come First

Supanova's execution model follows a strict cascade:

Objective → Project → Task → Subtask

An objective is a strategic goal for a quarter. Projects are the plans that achieve it. Tasks are the individual units of work agents execute. Subtasks are the granular steps within tasks.

If you skip OKRs and go straight to creating projects, the system still works — but your agents won't have strategic context. They'll execute tasks without understanding why they matter, which means they can't make smart tradeoffs when priorities conflict.

Setting OKRs first gives every downstream decision a frame of reference.

Step 1: The Question Universe (Your Strategic Foundation)

Before you write a single objective, Supanova needs to understand your business. That happens through the Question Universe.

During workspace setup, you went through a flow titled "Training Your AI Strategy Team." This is a set of 50 strategic questions across six categories:

Product & Service — what you sell, your value proposition, revenue model, customer pain points
Team & Culture — founder story, team size, strengths, hiring gaps, decision-making style
Growth & Innovation — 3-year vision, growth priorities, untapped opportunities, AI readiness
Market & Competition — ideal customer profile, acquisition channels, competitive landscape, churn reasons
Operations & Efficiency — daily operations, bottlenecks, manual processes, tech stack
Financial Health — revenue trends, margins, unit economics, funding needs, cash position

Each answer includes a confidence score you set (0-100%). The system uses these scores to weight how heavily it relies on each answer when making strategic decisions.

If you skipped some questions during setup, you can return to them anytime in Workspace Settings. The system also flags answers older than 90 days as potentially stale, prompting you to update them as your business evolves.

Why this matters for OKRs: When you generate objectives with AI, the platform reads your Question Universe answers to understand your market position, growth goals, team capabilities, and financial constraints. The richer your answers, the more relevant the generated objectives will be. Vague answers produce generic objectives. Detailed, high-confidence answers produce objectives that actually reflect your business.

Step 2: Navigate to the Strategy Page

Go to /strategy from the sidebar. This is where all your objectives live, organized by quarter (e.g., "Q2 2026").

You have two paths: generate objectives with AI or create them manually.

Step 3: Option A — Generate Objectives with AI

Click "Generate with AI." This activates the platform's strategic intelligence layer — the part of Supanova responsible for understanding your business context and translating it into actionable direction.

Here's what happens behind the scenes:

The platform pulls your Question Universe answers and evaluates confidence levels
It reviews your active projects, team composition, and current objectives
It reads your workspace's mission and industry context
It generates 3-5 draft objectives for your selected quarter

Each draft objective includes:

Title — a clear statement of what you're trying to achieve
Description — the strategic rationale: why this objective matters now
Key Results — 3-5 measurable outcomes, each with a target value and unit of measurement (e.g., "Reduce customer onboarding time to 48 hours")
Estimated projects — how many projects the system thinks it will take
Priority — low, medium, high, or critical

These drafts arrive in draft status. You review each one and either activate it (meaning the system starts acting on it) or archive it (saved but not pursued).

The key principle: AI proposes, you decide. The platform is opinionated about what your business should focus on — but you have final say.

Step 3: Option B — Create an Objective Manually

Click "Create Objective" and fill in:

Title: What you want to achieve. Be specific. "Grow revenue" is vague. "Launch self-serve checkout and reach 200 paying customers" gives agents something concrete to work toward.
Description: Why this matters. What does the business look like when this objective is complete? What changes?
Quarter: Which quarter to target. The system provides quarter options for the current and upcoming periods.
Priority: How urgent this is relative to your other objectives. Priority affects which projects get resources first.
Key Results: Add 3-5 measurable outcomes.

For key results, you can type them manually or click "Generate Key Results." The AI analyzes your title and description and suggests KRs with specific target values and units — like "Increase monthly active users" with a target of 10,000 and a unit of "users." You can edit, keep, or discard each suggestion.

You can also click "Refine with AI" after creating an objective. The platform reviews your text and returns an improved version with 3 specific suggestions for better clarity and strategic alignment.

How Key Results Track Progress

Each Key Result has:

Title: The measurable outcome (e.g., "Reduce customer churn rate")
Target value: The number you're aiming for (e.g., 5)
Unit: What you're measuring (e.g., "%")
Status: on_track, at_risk, behind, or completed

As agents complete tasks in projects linked to the objective, progress rolls up. You don't need to manually update KR status — the system calculates it from actual work completed.

What Happens After You Activate an Objective

Activating an objective sets a cascade in motion:

The platform marks the objective as active and signals the governance layer
Supanova evaluates the objective and begins creating projects — breaking it into concrete execution plans with timelines, deliverables, and task breakdowns
The system reviews the required skills for generated projects and assigns agents from your workforce (or recommends hiring new ones)
Agents begin executing — autonomously, or with your approval at each step, depending on your execution mode setting

You can also link existing projects to objectives manually from the objective detail view.

The Objective Lifecycle

Objectives move through four statuses:

Draft: Generated by AI, waiting for your review. No downstream action happens.
Active: You've approved it. The platform is actively creating projects and assigning resources.
Completed: The quarter ended and the objective was achieved. Success patterns are extracted.
Archived: You chose not to pursue it, or it was deprioritized. Saved for reference.

How Your Objectives Train the Workforce

This is where the compounding value kicks in.

When an objective completes successfully, the system extracts what worked — what kinds of tasks succeeded, what agent configurations delivered, what strategic context was most useful. These success patterns are stored with high confidence and referenced in future decisions.

When an objective expires without completing, the system runs failure analysis — what blocked progress, which assumptions were wrong, where resources were misallocated. These anti-patterns are also stored, so the same mistakes aren't repeated next quarter.

Over time, this means:

AI-generated objectives get more aligned with what actually works for your business
Project breakdowns become more realistic based on historical execution data
Agent assignments improve as the system learns which skill combinations succeed

Every objective you set — whether it succeeds or fails — is training data for a smarter workforce next quarter. The system doesn't just execute your strategy. It learns from it.

Tips for Your First OKRs

Answer your Question Universe thoroughly first. The AI generates objectives from your answers. Sparse answers produce generic objectives.
Start with 3-5 objectives per quarter. More than that spreads your AI workforce too thin. Focus beats volume.
Use priority honestly. Critical means this quarter fails without it. High means it's important but not existential. The priority field affects resource allocation across every project and task.
Review AI drafts carefully before activating. Once active, the system starts creating projects and assigning agents. Make sure the objective actually reflects what you want.
Set confidence scores on your Question Universe answers. A 90% confidence answer carries more weight than a 40% answer. If you're guessing, say so — the system adjusts.
Update stale answers. If it's been 90+ days since you answered a strategic question and your business has changed, update it. The next round of AI-generated objectives will be better for it.

What's Next

Your objectives are set. The platform is creating projects. In the next guide, we'll walk through what those projects look like and how to submit your own: Creating Your First Project.

... Six