A little perspective.
A clearer next move.

Start with one question.

Orlog brings customer groups, product directions, evidence, and decisions into one workspace. Product–market fit is one use case; pricing, market entry, competitive positioning, and organizational choices belong here too.

1. Choose a customer and direction.Open the strategy matrix and select a cell that represents a decision you want to examine.
2. Look through several lenses.Compare fit, revenue, friction, and positioning. Inspect the source context before trusting a number.
3. Test the assumptions.Try a simulation, define an experiment, and record what would change your mind.

One decision. Several perspectives.

Rows represent customer groups. Columns represent product directions. A cell connects the two. Select it to see the hypothesis, linked evidence, and a suggested next experiment.

Present the story

Open the guided presentation for a four-step walkthrough: understand the board, start with the customer, compare revenue, and identify the next experiment. Use Previous and Next or the step buttons. Auto-play advances the story; Pause gives you time to discuss. Close or Escape returns to the workspace.

The layer stack is a way to compare views of the same choice. It is not an additional prediction model. SWOT uses a separate, human-authored worksheet.

How scoring works.

Demo scores start from fictional priors on a 1–5 scale. They help explain the interface and are not validated forecasts, probabilities, or recommended investment decisions. Private workspaces start unscored.

Evidence counts are not confidence estimates. A low number of sources, contradictory observations, or an untested prior is a reason to investigate further. Automatic evidence-backed scoring and calibration are not available in the private pilot.

Evidence has a source.

The evidence library holds observations and statements alongside their source references. Search the library, inspect a record, and see which customer group and strategic direction it informs.

In a configured private workspace

1. Import plain text, Markdown, or CSV notes, up to 50 KB per file. Provide a title and source reference.
2. Choose Analyze with AI to send the source text to the configured OpenAI provider for embeddings and proposed claims.
3. Review each proposal against its exact source quote. Approve or reject it, then assign a persona, direction, and supporting, contradicting, or neutral relationship.
4. Use approved evidence to inform your assessment. An approved quote does not automatically prove the claim is true.

Importing stores the document; analysis is a separate action. Removing a document also removes its derived chunks and claims. The public workspace does not accept uploads or change its sample evidence.

Research from what you know.

Start with a specific question: “What blocks adoption for engineering teams?” or “What evidence supports an API-first direction?” Inspect returned sources and distinguish what they say from what you infer.

Exact-quote checks and citation-ID validation reduce unsupported output; they do not independently verify every conclusion. Background deep research and autonomous web research are planned, not currently running.

Personas are working hypotheses.

Open a persona to understand the customer group’s needs, decision context, barriers, and next questions. Use linked evidence to challenge the profile, then explore its row in the matrix.

The demo profiles are authored examples. The private pilot supports manually created personas and needs. Automatic segmentation, discovered organizational structures, and AI-generated persona profiles are future capabilities.

Move the assumptions. Keep the baseline.

The sensitivity model lets you vary price, adoption, and operating costs independently. Load a scenario’s assumptions or adjust the sliders. Reset to baseline returns all three changes to zero.

Percent changes are expressed as decimals in these formulas: 10% becomes 0.10. The baseline has revenue of 100, operating costs of 60, and contribution of 40. All three displayed indices equal 100 at baseline.

There is no assumed price elasticity, causal chain, uncertainty distribution, or competitive response model. Negative contribution values represent losses relative to the baseline contribution. Demo exploration does not save shared branches. A configured private workspace can save named assumptions and track the corresponding hypothesis in the decision log.

Remember why you chose.

The decision log brings the choice, expected outcome, success metric, and review date together. The example decision in the demo shows the structure; it is not an actual company decision.

1. Write a concrete decision and the outcome you expect.
2. Specify a measurable success criterion and a date to review it.
3. After the experiment, record what happened, including measured results.
4. Compare the outcome with the original hypothesis and decide what to change.

Private owners can record decisions and outcomes. The current pilot does not automatically ingest business outcomes or calculate prediction accuracy.

Connect evidence deliberately.

Connection cards show where business context can come from. Demo connection states and records are examples; clicking a card does not authorize a provider or access an account.

Provider adapters require operator configuration and a separately authorized account. Live customer connections have not been validated. Sync imports source documents; it does not automatically run paid AI analysis. Disconnect removes local connection data and imported evidence, but does not revoke the grant at the provider. Continuous synchronization and team permission propagation are not implemented.

People, workspace, and preferences.

Open the workspace switcher or profile to inspect Northstar Labs, its planning focus, and sample team. Teams and Users show how product, engineering, and revenue perspectives can work together. Owner, Editor, and Viewer labels illustrate roles; they do not enable public editing.

Settings collects workspace details, access information, and a reduced-motion preference. The in-app motion preference lasts for the current session; the system’s reduced-motion preference is also respected. Docs stays available from the sidebar. In the public demo, Sign out takes you back to the homepage.

The private pilot is restricted to one explicitly configured workspace owner. Invitations, role management, additional workspaces, and general account creation are not yet available.

A working foundation, with clear boundaries.

Availability in a demo screen does not mean a live integration or production team workflow is enabled. The public demo cannot access private evidence or trigger paid AI calls.