Pilot Business OS
Help Center

Onboarding & documentation

Everything you need to connect Pilot to your ERP and turn data into executive decisions.

How to connect Odoo ERP

Connect your Odoo environment to Pilot Business OS safely.

  1. 1Create or choose a dedicated Odoo user for Pilot.
  2. 2Give this user read access only to the modules needed for analysis.
  3. 3Open the Pilot onboarding or company settings page.
  4. 4Enter your Odoo URL, database name, username, and access key/password.
  5. 5Test the connection before saving.
  6. 6After successful connection, Pilot will import available ERP data into its internal normalized layer.
  7. 7AI analysis starts only after the data is synced into Pilot's internal database.

Odoo credentials setup

Recommended setup for a secure Pilot connection.

  1. 1Use a separate Odoo user instead of a personal administrator account.
  2. 2Give read-only access where possible.
  3. 3Avoid sharing full admin credentials unless absolutely required.
  4. 4Store credentials only inside the secured Pilot connection settings.
  5. 5If the Odoo password or key changes, update the connection settings in Pilot.
  6. 6If access is removed in Odoo, Pilot sync will stop until credentials are updated.
  7. 7For production use, coordinate access with the company's ERP administrator.

Data synchronization

How Pilot imports and refreshes ERP data.

  1. 1Pilot connects to the approved ERP source configured for the company.
  2. 2ERP data is imported into Pilot's standardized internal data layer.
  3. 3AI modules work only on normalized Pilot data, not directly on live ERP screens.
  4. 4Manual refresh can update business intelligence, but it should not directly train ML models.
  5. 5Background refresh can keep dashboards and recommendations up to date.
  6. 6If ERP connection fails, the system should show a clear connection or sync status.
  7. 7Historical learning and business memory should be preserved according to Pilot's architecture.

AI recommendations

How Pilot turns ERP data into business actions.

  1. 1Pilot analyzes sales, inventory, procurement, and production signals where data is available.
  2. 2Recommendations are generated from internal normalized data and saved results.
  3. 3Each recommendation should have a business reason, evidence, and confidence level.
  4. 4The user should be able to understand why a recommendation was suggested.
  5. 5Unknown or insufficient data should be shown honestly instead of fake advice.
  6. 6Recommendations should support decision-making; they should not automatically change ERP data.
  7. 7The goal is to help business owners decide what needs attention today.

Forecasting

How Pilot uses business data to support planning.

  1. 1Forecasting uses available sales and operational history from the company's ERP data.
  2. 2Forecast quality depends on the amount and quality of historical data.
  3. 3Pilot can highlight demand changes, growth opportunities, and possible risks.
  4. 4Procurement and production planning should rely on sales demand signals where available.
  5. 5Forecasts should be explained in business language, not only as technical scores.
  6. 6Low-data situations should be marked clearly as limited confidence.
  7. 7Forecasting improves over time as more clean business data becomes available.

Executive Copilot

How the Pilot Assistant helps owners and managers.

  1. 1The Assistant is designed as a business executive copilot, not a general chatbot.
  2. 2It answers using Pilot's cached business intelligence, snapshots, and evidence.
  3. 3It should not invent facts when data is missing.
  4. 4It can explain risks, opportunities, changes, and recommended actions.
  5. 5A key use case is answering: 'What should I do today?'
  6. 6Answers should include situation, why it matters, recommended action, evidence, and confidence.
  7. 7The Assistant should guide decisions, while the human remains in control.

Security and privacy

How Pilot should protect company data.

  1. 1Pilot should use separate company environments and tenant-aware access rules.
  2. 2Users should only see the company data they are allowed to access.
  3. 3ERP credentials should be stored securely and never shown publicly.
  4. 4Pilot should avoid writing back to ERP unless an explicit future feature is approved.
  5. 5AI should work on Pilot's internal normalized data layer, not directly on uncontrolled ERP data.
  6. 6Company-specific data should remain tenant-scoped.
  7. 7Any global learning layer should use only anonymized and aggregated patterns, not raw tenant data.